Introduction

National parks, as invaluable repositories of natural and cultural heritage, bear the weighty responsibility of ecological preservation while simultaneously serving as vital platforms for public relaxation, education, and travel1,2. Within these protected natural spaces, comfort emerges as a decisive component of visitor satisfaction, profoundly influencing their perception and quality of experience. Comfort, a broad concept, essentially equates to individual well-being in relation to their environment and also represents a state of balance between the person and their surroundings3. As posited by McNally et al.4comfort encompasses physical, physiological, and psychological elements. While the physical and physiological aspects, tied to environmental conditions, can be more objectively assessed and impact all individuals, psychological aspect is highly subjective, varying according to individuals’ sociodemographic profiles and personality traits. The experience of comfort is further defined as a harmonious state of happiness, achieved through the coordination and consistency between an individual’s physical, physiological, and psychological aspects and their environment5. This holistic comfort experience is shaped by a multitude of factors, including physical elements that contribute to a amenity space, such as lighting, noise levels, and ventilation, as well as psychological factors that foster satisfaction in both social and natural environments.

In the context of national parks, recreation is viewed as a transformative experience, profoundly influencing individuals’ internal changes6. During these experiences, tourists often engage in reflection and re-evaluation of their perspectives, attitudes, lifestyles, or behaviors7. The impact of transformative experiences is instrumental in fostering personal growth, societal development, and cultural exchange, playing a pivotal role in shaping both individuals and societies7. Empirical research has demonstrated that natural environments provide comfortable experiences, alleviating negative emotions and stress, with a significant correlation between pursuing comfort and enhancing life satisfaction and happiness5,8. Post-COVID-19, a shift in tourist preferences has been observed, with many seeking the spiritual comfort in the expansiveness of public spaces9. Furthermore, ensuring comfort experiences and positive states for tourists may increase the likelihood of transmitting these positive states to others10. From the perspective of tourists’ perceived transformative experience, comfort experience of recreation in national parks transcends a mere physical, physiological, or psychological state. It is seen as a continuous process of maintaining optimal recreational states, leading to sustained improvement and enhancement of tourists’ transformative experiences. Therefore, comfort experience of recreation should be reconceptualized to encompass a broader range of dimensions and facets.

Most studies about comfort in recreation can be grouped into two themes: focusing on the objective comforts by the supply side and the tourists’ subjective recreation experience by the demand side. The first group examined the tangible aspects that contribute to a tourist’s comfort level, such as the availability and quality of facilities and amenity services5,11. This includes environmental factors like natural light influx, eye-strain minimizing lighting, noise control, and efficient ventilation systems, all crucial in creating a physically and mentally soothing atmosphere. Conversely, the second group was attempting to understand and characterize these experiences through personal narratives. However, existing studies often limit their analysis to sensory experiences, neglecting the broader spectrum of comfort needs12. From a psychological perspective, comfort is not just a state but a dynamic process involving cognition, memory, and expectation, with cultural attitudes toward nature also shaping comfort perceptions2. Additionally, the concept of social comfort emerges, suggesting that comfort perceptions are influenced not only by individual preferences but also by group norms and standards. Given these insights, the necessity for an integrative comfort evaluation model in recreation to comprehensively address the multifaceted nature of the transformative experience in national parks13.

Based on the aforementioned analysis, we take Wuyishan National Park as an example to introduce the emerging concept of recreation comforting degree that integrates comforting experience with direct nature and amenities contact to enhance tourists’ perceived transformative experience. This paper aims to contribute to a deeper understanding of tourists’ genuine perceptions of comforting degree provided by National Park recreation activities and to assist National Park managers in optimizing key factors influencing recreation comforting degree. To this end, we first propose the recreation comforting degree in National Parks measurement tool, using a mixed-methods that combines Text Mining (TM) technology with the fuzzy DEMATEL-based analytic network process (FDANP). Previous literature typically summarized factors through literature reviews and subsequently determined final factors through in-depth interviews with experts14. While this method leverages expert experience to establish evaluation indicators, it is subject to criticism regarding subjectivity and uncertainty. To address this gap, our study employs TM techniques to extract online reviews from online travel agencies (OTAs) and analyzes high-frequency words to identify influential factors affecting recreation comforting degree. Subsequently, through a three-level coding process based on grounded theory and expert focus group discussions, the high-frequency words are categorized and named. Then, the fuzzy DEMATEL and DANP methods are used to explore the causal relationships between the factors and their relative importance. The resulting of Influential Relationship Map (IRM) provides directions for improving National Park recreation comforting degree, thereby rendering improvement strategies more effective. This research extends the existing literature on comfort by focusing on tourists’ perspectives on the dimensions of recreation comforting degree within the context of tourism and hospitality in National Parks.

Literature review

The concept of comfort experience has been explored across various disciplines, with ongoing debate in the literature regarding its definition15. The first perspective identifies only two states: discomfort present and discomfort absent, introducing comfort as a neutral state that does not necessarily entail positive effects such as pleasure16. It suggests that comfort, as a physical phenomenon, may not truly exist17. The second viewpoint describes optimal comfort conditions as the physiological state of the body in relation to the surrounding environment18. The physical aspect emphasizes sensory perceptions, including temperature, humidity, and air quality19,20. Supporting this perspective, recent research demonstrates that forest environments, through factors such as canopy structure, plant coverage, and microclimate conditions, can significantly shape human physiological comfort21. Building on this, comfort has been extensively studied as a crucial attribute in the realm of guest satisfaction, with the psychological aspect encompassing mood, stress, and satisfaction22. In shopping malls, comfort is characterized by the influx of natural light, while lighting, noise, and ventilation facilities are essential for creating an atmosphere conducive to both physical and mental comfort23. Aircraft passengers define comfort as a pleasant state of well-being, ease, and harmony between a person and their environment15. However, in reality, there is a distinction between comfort experience and pleasure. The feelings of comfort and discomfort are related to the level of arousal and depend on whether arousal is at its optimum level24. Excessive comfort may preclude pleasure, therefore, individuals sometimes must choose between pleasure at the expense of some comfort and more complete comfort at the sacrifice of pleasure. The final perspective posits that the comfort experience is continuously constructed by society. For instance, on beaches, comfort experience is understood as the level of satisfaction of beach goers concerning social variables25. Social comfort is influenced by interactions with local populations and other travelers26. In the field of national park recreation research, previous studies have primarily focused on the pleasure and satisfaction derived from recreational activities27.

To date, there has been limited research on comfort in the context of recreation. Particularly, spiritual comfort, akin to the concept of ningjing or jing hua xin ling is increasingly recognized as an integral part of the natural park experience28. The research demonstrates that exposure to nature can enhance the meaning of life by increasing savoring, thereby further highlighting the significance of natural environments for psychological comfort and spiritual solace29.These descriptions of comfort demonstrated its complexity, yet they all have their limitations. They consider only certain aspects of comfort while disregarding that comfort is a multidimensional construct encompassing physical, physiological, and psychological comfort, among others5. It is also specific to a particular time, situation, and individual. Therefore, it can be argued that the failure to understand this complexity has led to previous studies of recreation comforting degree in national park being based on limited definitions of comfort13. This study introduces a more specific type of comfort in national parks, namely recreation comforting degree, as a holistic state of being that harmonizes physiological, psychological, spiritual, and social comforting degree30 enhancing tourists’ perceived transformative experience, thereby contributing to the sustainable development of national park tourism.

Numerous scholars have employed questionnaires as a means to collect user comfort data, utilizing statistical methods to quantify comfort levels across diverse populations and analyze the pivotal factors influencing user comfort31. Richards32 developed a continuous scale for assessing various degrees of passenger comfort. Guest comfort was conceptualized as an internal state influenced by stimuli encompassing both physical and psychological comfort, leading to the development of an 11-indicator measurement scale for guest comfort in hotel settings33. In contrast, two distinct scales were devised to express users’ perceived chair comfort and discomfort, respectively34. Additionally, content analysis of written reports on comfort and discomfort experiences or a compilation of comfort descriptors represents another methodology employed for developing tools (e.g., questionnaires) to evaluate comfort experiences15. Qualitative methodologies provide insights into the subjective experiences and comfort needs of tourists. Depth interviews and participant observation are crucial in unveiling the nuanced perceptions and personal narratives that delineate the comfort experience35.

In the digital era, social media and travel review platforms are teeming with User Generated Content (UGC), encompassing a wide array of comments, ratings, photos, and videos. This UGC offers an exceptionally suitable source of accessible, reliable, credible, and readily available information for tourists and service providers36. Recent research has underscored the significance of web-generated content in comfort studies. For instance, a novel measurement tool employing a mixed-methods has been developed by combining qualitative narrative analysis of web reviews written by glamping tourists with a survey of glamping guests37. By mining online reviews and social media content, researchers can extract and quantify comfort-related themes, providing a holistic view of tourist comfort that encompasses both objective measurements and subjective experiences13. This is instrumental in discerning nuanced preferences and comfort-related needs of tourists and to construct dimensions of evaluation indicators through text mining algorithms. Furthermore, the Decision Making Trial and Evaluation Laboratory (DEMATEL) method excels in dissecting complex issues by elucidating the intricate web of relationships among various factors. In the tourism domain, DEMATEL facilitates a granular analysis of destination attributes, including the natural environment, cultural heritage, facilities, and accessibility, thereby aiding strategic tourism planning and development38. Fuzzy DEMATEL further aids in analyzing the interrelationships between uncertainty and factors within fuzzy settings, enabling a more targeted approach to enhancing recreational experiences. By integrating text mining with Fuzzy DEMATEL, we construct a model that not only identifies but also quantifies the interdependencies among comfort factors. This model serves as a cornerstone for decision-makers in National Parks to formulate strategies that align with tourists’ comfort experiences and represents a helpful strategy for improving tourist comfort39.

Methods

The analytical process is segmented into three primary components (Fig. 1). In the initial phase, Python 3.10 web scraping techniques are employed to gather online reviews from OTAs. Subsequently, the R language’s “jieba” word segmentation package is utilized for textual data cleaning, followed by the computation of the Term Frequency-Inverse Document Frequency (TF-IDF) value for each feature word. The second component involves constructing dimensions and indicators based on high-frequency vocabulary. First, high-frequency vocabulary is organized and classified through a three-level coding process based on grounded theory. Subsequently, based on a review of relevant literature and focus group expert discussions, dimensions and sub-dimensions are named to construct a recreation comforting model. In the third stage, the impact of recreation comforting indicators is assessed based on the developed indicator system and expert judgment.

To this end, the fuzzy DEMATEL method is used to numerically process the initial equilibrium matrix. This paper employs the CFCS method proposed by Wang and Tzeng40 to convert fuzzy data into crisp scores, obtaining the initial impact matrix for each expert through the fuzzy aggregation process. Next, the total impact matrix of fuzzy DEMATEL is calculated, and a network relationship diagram is generated. Then, the fuzzy DANP method is applied, in conjunction with the principles of analytic network process. The total impact matrix is normalized and transposed (ensuring that the sum of each row equals 1), and multiple matrix multiplications are performed to obtain the final total impact matrix of fuzzy DEMATEL, thereby refining the importance ranking of the indicators. The study was approved by the local ethics committee of Wuyi University. The study adhered strictly to Scientific Reports guidelines and the Declaration of Helsinki for human participant research, excluding transplantation studies and vulnerable groups’ involvement without adequate consent. Written informed consent was secured from all participants before expert interviews. Gender and sex equity principles were upheld, with clear distinction and consideration of both terms in research design and analysis.

Fig. 1
figure 1

Experimental process.

Full size image

Text mining and content analysis

OTAs, serving as intermediaries between tourism providers and consumers, have amassed a plethora of user reviews and travel narratives8. These accounts offer an authentic lens into tourists’ experiences and emotional responses. The feedback collected on OTAs holds paramount significance, as it guides potential customers and directly impacts service improvements41. Potential consumers generally perceive tourist reviews from OTAs as more credible than promotional materials from tourism providers42,43. A comprehensive text-mining analysis of these reviews and travel narratives is pivotal in unveiling tourists’ genuine experiences and enhancing recreation comforting degree. Text mining techniques excel at extracting valuable insights from vast textual data. Extensive scholarly research has leveraged text mining to analyze online travel reviews, employing methods such as sentiment analysis and topic modeling to assess tourist satisfaction, uncover their needs44 and identify challenges with products and services45. This information is indispensable for tourism providers seeking to refine their market positioning and enhance their competitive edge46,47.

In this study, we collected data from OTAs and yielded a dataset of online reviews and comprehensive travel narratives. To enhance the analytical utility of the dataset, a rigorous preprocessing approach was implemented, encompassing: (1) exclusion of reviews lacking substantive content; (2) removal of emoticons and non-English text; (3) concise extraction of pertinent information from lengthy reviews; and (4) standardization of textual data through spelling and grammatical corrections, as well as synonym harmonization. This meticulous process culminated in a refined dataset of reviews and comprehensive travel narratives. Subsequently, content analysis allows researchers to explore texts independently, without the influence of external theories or concepts. In this process, a combination of the three-level coding procedure of grounded theory and expert discussions in focus groups was used. Three researchers collaboratively conducted iterative coding and labeling work. The participants’ focal points were extracted through these labels and grouped and classified based on similarity. Ultimately, items describing recreation comforting degree were identified and categorized through a triangulation method involving vocabulary and literature. Additionally, the expert group engaged in in-depth discussions and naming of these items and their dimensions to ensure the validity of the content44. This method effectively alleviated the limitations of a single source of information, allowing for the inclusion and cross-validation of diverse perspectives, thereby enhancing the validity of the research45. The expert group consisted of two managers from Wuyishan National Park and an academic researcher in the field of leisure.

Fuzzy DEMATEL and DANP

After confirming the recreation comforting degree, we employ Fuzzy DEMATEL to explore the network relationships among indicators. An expert questionnaire based on DEMATEL was administered to 12 tourism professionals, including 4 tourism managers, 6 tourism scholars, and 2 travel enthusiasts. The framework was structured into two stages: (1) constructing an Impact-Relations-Map (IRM) with Fuzzy DEMATEL among indicators, and (2) calculating the influence weight of each criterion using DANP48.

Fuzzy DEMATEL for constructing IRM

This study follows the research of Kazancoglu and Aksoy to perform the steps of fuzzification and defuzzification, processing the DEMATEL questionnaire data from each expert to obtain the fuzzy initial influence matrix for each expert49.The triangular Fuzzy number and its corresponding fuzzy number were defined in (Table 1). The steps of the Fuzzy DEMATEL method are given below.

Table 1 Linguistic variable and triangular fuzzy number.
Full size table

Step 1: Pooling the experts’ ratings, the average fuzzy initial direct relationship matrix \({\varvec{A}}\) is calculated.

Let \({\varvec{a}}_{{ij}}^{H}=({\varvec{a}}l_{{ij}}^{k},{\varvec{a}}m_{{ij}}^{k},{\varvec{a}}r_{{ij}}^{k})\), having k experts, the fuzzy weight of \({\varvec{a}}{\varvec{m}}_{{{\varvec{i}}{\varvec{j}}}}^{k}\) denotes the \({i^{th}}\) criteria, and it affects the \({j^{th}}\) criteria to be evaluated by \({k^{th}}\) experts.

(1) Normalization:

$$\begin{array}{*{20}{c}} {\quad \varvec{\varDelta}_{{\hbox{min} }}^{{\hbox{max} }}=\hbox{max} ar_{{ij}}^{k} - \hbox{min} al_{{ij}}^{k}\quad \quad \;\;\;} \\ {xal_{{ij}}^{k}=(al_{{ij}}^{k} - \hbox{min} al_{{ij}}^{k})/\varvec{\varDelta}_{{\hbox{min} }}^{{\hbox{max} }}} \\ {{\varvec{x}}am_{{ij}}^{k}=(am_{{ij}}^{k} - \hbox{min} al_{{ij}}^{k})/\varvec{\varDelta}_{{\hbox{min} }}^{{\hbox{max} }}} \\ {{\varvec{x}}ar_{{ij}}^{k}=(ar_{{ij}}^{k} - \hbox{min} al_{{ij}}^{k})/\varvec{\varDelta}_{{\hbox{min} }}^{{\hbox{max} }}} \end{array}$$
(1)

(2) Calculate the normalized values of left (ls) and right (rs):

$$\begin{array}{*{20}{c}} {{\varvec{x}}{\varvec{l}}{\varvec{s}}_{j}^{k}={{{\varvec{x}}am_{{ij}}^{k}} \mathord{\left/ {\vphantom {{{\varvec{x}}am_{{ij}}^{k}} {(1+{\varvec{x}}am_{{ij}}^{k} - {\varvec{x}}al_{{ij}}^{k})}}} \right. \kern-0pt} {(1+{\varvec{x}}am_{{ij}}^{k} - {\varvec{x}}al_{{ij}}^{k})}}} \\ {{\varvec{x}}{\varvec{r}}{\varvec{s}}_{j}^{k}={{{\varvec{x}}ar_{{ij}}^{k}} \mathord{\left/ {\vphantom {{{\varvec{x}}ar_{{ij}}^{k}} {(1+{\varvec{x}}ar_{{ij}}^{k} - {\varvec{x}}am_{{ij}}^{k})}}} \right. \kern-0pt} {(1+{\varvec{x}}ar_{{ij}}^{k} - {\varvec{x}}am_{{ij}}^{k})}}} \end{array}$$
(2)

(3) Calculate the overall normalized crisp value:

$${\varvec{x}}_{j}^{{cris{p^k}}}={{[{\varvec{x}}{\varvec{l}}{\varvec{s}}_{{ij}}^{k}(1 - {\varvec{x}}{\varvec{l}}{\varvec{s}}_{{ij}}^{k})+{\varvec{x}}{\varvec{r}}{\varvec{s}}_{{ij}}^{k} \cdot {\varvec{x}}{\varvec{r}}{\varvec{s}}_{{ij}}^{k}]} \mathord{\left/ {\vphantom {{[{\varvec{x}}{\varvec{l}}{\varvec{s}}_{{ij}}^{k}(1 - {\varvec{x}}{\varvec{l}}{\varvec{s}}_{{ij}}^{k})+{\varvec{x}}{\varvec{r}}{\varvec{s}}_{{ij}}^{k} \cdot {\varvec{x}}{\varvec{r}}{\varvec{s}}_{{ij}}^{k}]} {[1 - {\varvec{x}}{\varvec{l}}{\varvec{s}}_{{ij}}^{k}}}} \right. \kern-0pt} {[1 - {\varvec{x}}{\varvec{l}}{\varvec{s}}_{{ij}}^{k}}}+{\varvec{x}}{\varvec{r}}{\varvec{s}}_{{ij}}^{k}]$$
(3)

(4) Calculation of crisp values for each expert questionnaire:

$$a_{{ij}}^{k}=\hbox{min} al_{{ij}}^{k}+{\varvec{x}}_{j}^{{cris{p^k}}} \cdot \varvec{\varDelta}_{{\hbox{min} }}^{{\hbox{max} }}$$
(4)

(5) Integration of crisp values from all expert questionnaires:

$${\overline {{\varvec{a}}} _{ij}}=\frac{1}{k}({\varvec{a}}_{{ij}}^{1}+{\varvec{a}}_{{ij}}^{2}+ \cdots +{\varvec{a}}_{{ij}}^{k})$$
(5)

The value of the fuzzy direct correlation matrix \({\varvec{A}}={[{\text{ }}{\overline {{\varvec{a}}} _{ij}}]_{n \times n}}\) is derived from the average of the same criteria in the various fuzzy direct matrices via Eq. (1)-Eq. (5) for all questionnaire respondents.

Step 2: Calculating the Normalized Average Direct-Influence Relation Matrix.

Normalize the mean matrix \({\varvec{A}}\) to obtain the initial impact matrix \({\varvec{E}}={[{\text{ }}{{\varvec{e}}_{ij}}]_{n \times n}}\), that is, the matrix\({\varvec{E}}\)can be obtained through Eq. (6), in which all principal diagonal elements equal zero.

$${\varvec{E}}={{\varvec{A}} \mathord{\left/ {\vphantom {{\varvec{A}} {\hbox{max} }}} \right. \kern-0pt} {\hbox{max} }}\{ \mathop {\hbox{max} }\limits_{{\text{i}}} \sum\limits_{{{\text{j=1}}}}^{{\text{n}}} {\left| {{{\overline {{\varvec{a}}} }_{ij}}} \right|} \;,\;\mathop {\hbox{max} }\limits_{j} \sum\limits_{{{\text{i=1}}}}^{{\text{n}}} {\left| {{{\overline {{\varvec{a}}} }_{ij}}} \right|} \}$$
(6)

Step 3: Matrix Total Influence-Relationship Matrix.

Based on the direct relation matrix, the total influence matrix is calculated.

$${\varvec{T}}={[{\text{ }}{{\varvec{t}}_{ij}}]_{n \times n}}={\varvec{E}}+{{\varvec{E}}^{\text{2}}}+{{\varvec{E}}^{\text{3}}}+ \cdots \approx {\varvec{E}}{(I - {\varvec{E}})^{ - {\text{1}}}}$$
(7)

Step 4: Constructing the Impact Network Relationship Map (IRM)

$${\varvec{r}}={[{\text{ }}{{\varvec{r}}_{ij}}]_{n \times {\text{1}}}}={\left[ {\sum\limits_{{{\text{j}}={\text{1}}}}^{{\text{n}}} {{{\varvec{t}}_{{\text{ij}}}}} } \right]_{{\text{n}} \times {\text{1}}}}\,,\,\,{\varvec{c}}={[{\text{ }}{{\varvec{c}}_{ij}}]_{n \times {\text{1}}}}=\left[ {\sum\limits_{{{\text{i}}={\text{1}}}}^{{\text{n}}} {{{\varvec{t}}_{{\text{ij}}}}} } \right]_{{{\text{1}} \times {\text{n}}}}^{{\text{T}}}\,$$
(8)

Where \({\varvec{r}}\) value represents the extent to which an indicator influences other indicators, while \({\varvec{c}}\) value reflects the extent to which the indicator is influenced by other indicators. Subsequently, and \({\varvec{r}}{\text{+}}{\varvec{c}}\) are calculated, where \({\varvec{r}}{\text{+}}{\varvec{c}}\) is known as the centrality degree, which indicates the importance of the indicator in multiple problems, while \({\varvec{r}}{\text{-}}{\varvec{c}}\) is known as the cause degree, which indicates the causal attribution of the indicator in multiple problems. When the \({\varvec{r}}{\text{-}}{\varvec{c}}\) value is positive, it indicates that the indicator tends to act as a cause; conversely, it indicates that it tends to act as an effect.

Calculating the influence weight of each criterion using DANP

The DEMATEL-based ANP (DANP) allocates weights to diverse criteria via a thorough evaluation system. This approach takes into account the interplay among clusters, defines these interactions through the Fuzzy DEMATEL questionnaire, and forms a networked hierarchy to determine the significance of evaluation factors. The steps encompassed in this process are as follows:

Step 5: Establish the Unweighted Supermatrix.

By using the total impact matrix \({\varvec{T}}_{{}}^{\varvec{\alpha}}\) and the threshold \(\varvec{\rho}\) being the average of all elements of \({\varvec{T}}\), a new matrix is generated. If the elements in matrix \({\varvec{T}}_{{}}^{\varvec{\alpha}}\) have values less than \(\varvec{\rho}\), their value is reset to zero. The new \(\varvec{\rho}\)-cut total-impact matrix \({\varvec{T}}_{{}}^{\varvec{\alpha}}\), as Eq. (9).

$$\begin{array}{*{20}c} {\user2{T}^{\alpha } } & {\begin{array}{*{20}c} = & {\left[ {\begin{array}{*{20}c} {t_{{11}}^{\alpha } } & \cdots & {t_{{1j}}^{\alpha } } & \cdots & {t_{{1n}}^{\alpha } } \\ \vdots & {} & \vdots & {} & \vdots \\ {t_{{i1}}^{\alpha } } & \cdots & {t_{{ij}}^{\alpha } } & \cdots & {t_{{in}}^{\alpha } } \\ \vdots & {} & \vdots & {} & \vdots \\ {t_{{n1}}^{\alpha } } & \cdots & {t_{{nj}}^{\alpha } } & \cdots & {t_{{{}_{{nn}}}}^{\alpha } } \\ \end{array} } \right]} \\ \end{array} } \\ \end{array} \begin{array}{*{20}c} \to \\ {} \\ \to \\ {} \\ \to \\ \end{array} \begin{array}{*{20}c} {w_{1} = \sum\nolimits_{{j = 1}}^{n} {t_{{1j}}^{\alpha } } } \\ \vdots \\ {w_{i} = \sum\nolimits_{{j = 1}}^{n} {t_{{ij}}^{\alpha } } } \\ \vdots \\ {w_{n} = \sum\nolimits_{{j = 1}}^{n} {t_{{nj}}^{\alpha } } } \\ \end{array}$$
(9)

The \(\varvec{\rho}\)-cut total-impact matrix \({\varvec{T}}_{{}}^{\varvec{\alpha}}\) is to be regularized by dividing by \({{\varvec{w}}_i}\), as \({\varvec{T}}_{C}^{\varvec{\alpha}}\).

$${\varvec{T}}_{C}^{\varvec{\alpha}}=\left[ {\begin{array}{*{20}{c}} {{{{\varvec{t}}_{{11}}^{{}}} \mathord{\left/ {\vphantom {{{\varvec{t}}_{{11}}^{{}}} {{{\varvec{w}}_1}}}} \right. \kern-0pt} {{{\varvec{w}}_1}}}}& \cdots &{{{{\varvec{t}}_{{1j}}^{{}}} \mathord{\left/ {\vphantom {{{\varvec{t}}_{{1j}}^{{}}} {{{\varvec{w}}_1}}}} \right. \kern-0pt} {{{\varvec{w}}_1}}}}& \cdots &{{{{\varvec{t}}_{{1n}}^{{}}} \mathord{\left/ {\vphantom {{{\varvec{t}}_{{1n}}^{{}}} {{{\varvec{w}}_1}}}} \right. \kern-0pt} {{{\varvec{w}}_1}}}} \\ \vdots &{}& \vdots &{}& \vdots \\ {{{{\varvec{t}}_{{i1}}^{{}}} \mathord{\left/ {\vphantom {{{\varvec{t}}_{{i1}}^{{}}} {{{\varvec{w}}_i}}}} \right. \kern-0pt} {{{\varvec{w}}_i}}}}& \cdots &{{{{\varvec{t}}_{{ij}}^{{}}} \mathord{\left/ {\vphantom {{{\varvec{t}}_{{ij}}^{{}}} {{{\varvec{w}}_i}}}} \right. \kern-0pt} {{{\varvec{w}}_i}}}}& \cdots &{{{{\varvec{t}}_{{in}}^{{}}} \mathord{\left/ {\vphantom {{{\varvec{t}}_{{in}}^{{}}} {{{\varvec{w}}_i}}}} \right. \kern-0pt} {{{\varvec{w}}_i}}}} \\ \vdots &{}& \vdots &{}& \vdots \\ {{{{\varvec{t}}_{{n1}}^{{}}} \mathord{\left/ {\vphantom {{{\varvec{t}}_{{n1}}^{{}}} {{{\varvec{w}}_n}}}} \right. \kern-0pt} {{{\varvec{w}}_n}}}}& \cdots &{{{{\varvec{t}}_{{nj}}^{{}}} \mathord{\left/ {\vphantom {{{\varvec{t}}_{{nj}}^{{}}} {{{\varvec{w}}_n}}}} \right. \kern-0pt} {{{\varvec{w}}_n}}}}& \cdots &{{{{\varvec{t}}_{{nn}}^{{}}} \mathord{\left/ {\vphantom {{{\varvec{t}}_{{nn}}^{{}}} {{{\varvec{w}}_n}}}} \right. \kern-0pt} {{{\varvec{w}}_n}}}} \end{array}} \right]$$
(10)

Transposing this matrix yields the unweighted matrix \({{\varvec{W}}^\alpha }\).

$${{\varvec{W}}^\alpha }={\left[ {{\varvec{T}}_{C}^{\alpha }} \right]^T}$$
(11)

Step 6: Deriving Local Weights for Dimensions and Criteria.

By repeatedly multiplying the transposed total influence relation matrix until it converges to a long-term equilibrium value, the weights of each evaluation dimension and criterion, i.e., the DANP influence weights, are determined.

$$\mathop {\lim }\limits_{{k \to \infty }} {({{\varvec{W}}^\alpha })^k}$$
(12)

Step 7: Calculating Global Weights for All Criteria.

The overall weight for each criterion is obtained by multiplying its local weight with the corresponding dimension weight.

$${\varvec{w}}_{{criteria}}^{{global}}={\varvec{w}}_{{D\_c}}^{{local}} \times {\varvec{w}}_{D}^{{local}}$$
(13)

Results

This study selected Wuyishan National Park (WYNP) as the research subject (Fig. 2), located in the northern part of Fujian Province, China, centered approximately at 117°41′46″E and 27°43′34″N, covering an area of 1280 km². With a peak elevation of 2160.8 m and a substantial forest coverage rate of 87.86%, WYNP stands as a unique environmental treasure. In 1999, WYNP was inscribed on the UNESCO World Natural and Cultural Heritage list due to its exceptional natural and cultural values, and subsequently awarded the title of National Grade 5 A Tourist Attraction, highlighting its high-quality services, profound historical and cultural heritage, and universal appeal to visitors. According to 2023 data, the main scenic area of WYNP received 1.65 million tourists, with a total of 13.6814 million tourists visiting the Wuyishan city throughout the year, generating a total tourism revenue of 20.18 billion yuan, further consolidating its significant position in the provincial and national tourism markets. WYNP has accumulated extensive visitor review data on OTAs offering invaluable empirical foundations for in-depth analysis of recreation comforting degree (Appendix A). Therefore, WYNP serves as an excellent case study for exploring recreational comfort experiences and strategies for managing and developing tourism destinations.

Fig. 2
figure 2

Study area.

Full size image

*The map was generated by using ArcGIS Desktop 10.5 (https://www.esri.com/).

Text processing

Use Python 3.10 web crawling technology to collect online reviews from March 2021 to March 2024 from OTAs such as Ctrip, Tongcheng, Mafengwo, and Douyin. After deleting duplicate values and removing irrelevant comments and abnormal data, 1,692 online reviews and 50 comprehensive travel narratives had been retained, totaling 1.01 million words. The efficiency of the data was 90.91% and 86.20% respectively. Table 2 showed the frequency statistics of vocabulary that was analyzed to highlight the most significant themes within the discourse of recreation comforting degree in national parks.

Table 2 Frequency statistics of vocabulary.
Full size table

Based on text mining and vocabulary content analysis, we systematically categorized thematically similar vocabulary into distinct groups by analyzing their interrelationships and logical sequencing, as detailed in (Table 3). The evaluation framework of the recreation comforting degree was structured around four main criteria and thirteen sub-criteria. Figure 3 illustrated the weight ranking derived from OTA platform data, demonstrating the following descending order of main criteria: Psychological comforting degree (PSC), Social comforting degree (SOC), Spiritual comforting degree (SPC),  and Physiological comforting degree (PHC).

Table 3 Classification results (chinese terms with distinct meanings but identical English translations separate frequency counts in table).
Full size table
Fig. 3
figure 3

Weights of various criteria based on text exploration results.

Full size image

By reviewing the literatures, we have added three sub-criteria, including nurturing and conditioning one’s health (PHC3), addressing psychological issues (PSC4) and improving mental state (SPC2). Based on a triangulation of the item sources between expert and literature, 16 items were retained in the survey, with four items undergoing wording refinements for greater accuracy (two items show differences due to nuances in Chinese-English translation). At last, three new items (i.e., PSC3, SPC1, SPC3) were identified through content analysis, while three items (i.e., PHC3, PSC4, SPC2) were derived from the literature, and other ten items were mentioned in both the content analysis and the literature (Table 4).

Table 4 Criteria and definitions for measuring recreation comforting degree.
Full size table

Establishment of the influencing factor model

After extracting the essential criteria (factors) and establishing the evaluation framework (Tables 2 and 3) by text mining, the DANP method is applied to explore the cause and effect relationship between criteria. The questionnaire based on DEMATEL was sent to 12 tourism professionals. Through defuzzification and pairwise comparison, the initial matrix A (Table 5) was derived according to step 1, as described in Sect. 3. According to steps 2 and 3 of the DEMATEL procedure outlined in the previous section, the normalized matrix D (Table 6) and the total impact matrix T (Table 7) were obtained, respectively. The results demonstrated a high level of reliability, with consensus reached among all twelve experts. A high consistency ratio of 96.6% (slightly above 95%) and a consistency gap of 3.4% (slightly below 5%) provided strong support for the research findings.

Table 5 Initial matrix A.
Full size table
Table 6 Normalized matrix D.
Full size table
Table 7 Total influence matrix T.
Full size table

Then, the total influence and influenced degree was given and accepted by dimensions and standards (Table 8), according to steps 4. The highest r + c value of PSC indicated that after considering both direct and indirect effects, psychological comforting degree contributed the most to overall comfort. This was also confirmed by the r-c value of PSC (0.053), which was close to zero, indicating that the direct and indirect effects of PSC were almost equal, emphasizing the importance of enhancing psychological comforting degree when improving overall comfort. In addition, the r-c value of PSC was positive, indicating that PSC not only directly affected overall comfort but may also have indirectly improved overall comfort by affecting other dimensions.

Table 8 The total influence and influenced degree influence on dimensions and standards.
Full size table

The Influence Relationship Map (IRM) among these criteria was constructed. The four evaluation criteria were interconnected (Fig. 4), with PSC exhibiting the highest connectivity and exerting influence on the other three criteria. Changes in PSC significantly impacted the entire recreation comforting degree. In terms of causality, PHC had the highest r-c value, indicating that it affected the other three criteria.

Fig. 4
figure 4

The impact-relations-map within main-criteria of the recreation comforting degree.

Full size image

As shown in Fig. 5, PHC1 demonstrated the highest relevance, thus highlighting its central role among the sub-criteria of PHC. PHC3 was the most deeply affected indicator. Similarly, within the sub-criteria of PSC, PSC2 had the greatest overall impact, but with a negative r-c value, indicating that it was affected by other criteria. PSC1 had the second-highest overall impact and acted as a causal factor that influenced other criteria. For SPC, SPC2 had the most significant impact, but with a negative r-c value, suggesting it was influenced by other criteria. In contrast, SPC1 was a causal factor that affected other criteria. Finally, within SOC, SOC1 held the most important influence, but with a negative r-c value, indicating that it was influenced by other criteria.

Fig. 5
figure 5

The impact-relations-map within sub-criteria of the recreation comforting degree.

Full size image

The weights of recreation comforting degree

Using the DANP technique to weight each criterion (Table 9), following steps 5 to 7, it was found that PSC held the highest weight (0.268) among all evaluation criteria, emphasizing its primary role in enhancing the recreation comforting degree of national parks. SOC was the second most important criterion. The top five weighted criteria included PSC2, SOC1, PSC3, SPC4, and PSC1.

Table 9 Weights of each criterion.
Full size table

Discussion

The recreation comforting degree for transformative experiences

This study innovatively proposes PHC, PSC, SPC, and SOC as the four core dimensions constituting recreation comforting degree, thereby constructing a comprehensive framework of comfort experience. This framework is theoretically supported by transformative tourism, which often exhibits an inward-to-outward trajectory of transformation58. Tourists perceive destinations as stimuli for co-creating their experiences, potentially leading to multifaceted transformations, including physical, knowledge-based, and social changes59. Furthermore, different types of transformations are likely interconnected, and exploring this area can deepen our understanding of the transformation process and its outcomes. Notably, despite the rarity of incorporating SPC in tourism experience research, the recreation comforting degree conceptual framework developed in this study strongly supports its significance. In transformative experiences, tourists are often prompted to engage in self-exploration, thereby enriching their spiritual dimensions and seeking inner meaning. This introspective journey not only enhances their personal understanding but also fosters a deeper sense of meaning and fulfillment. Additionally, the sub-dimensions of the recreation comforting degree contribute to depicting tourists’ health and well-being, aligning with the philosophy of transformative experiences60. In conclusion, we argue that the recreation comforting degree model serves as a precursor to transformative experiences, holding vast research potential. The assessment of this model will provide valuable support for further advancing the enhancement of transformative experiences and informing decision-making related to health and well-being.

Interconnections among main-criteria of recreation comforting degree

The recreation comforting degree model proposed in this study, based on the DANP analysis, elucidates a comprehensive and interconnected framework comprising PHC, PSC, SPC, and SOC. This multidimensional system resonates with Rybczynski17 Onion Theory, which emphasizes the importance of comfort as a dynamic and context-specific construct, rather than a monolithic concept. Our model further contributes by delineating the interplay among these dimensions, highlighting their synergistic effect on optimizing the overall comfort experience (Fig. 6). In line with previous research, the present study found that PSC plays a pivotal role in the comfort system, with the highest r + c value indicating its substantial direct and indirect contributions to overall comfort. This finding echoes the assertion that psychological comforting degree is a critical component in enhancing overall well-being61. Specifically, our results demonstrate that PSC not only directly influences overall comfort but also indirectly enhances PHC, SPC, and SOC. This multifaceted influence aligns with Aggarwal et al.31,33 observation that sustainability practices in hotels significantly impact guests’ physical and psychological comforting degree, ultimately affecting their revisit and recommendation intentions. The direct effect of PSC on PHC observed in our study supports the notion that a favorable psychological state can foster physical well-being. This relationship is consistent with the biopsychosocial model, which posits that psychological factors can significantly impact physical health outcomes62. Similarly, the indirect effect of PSC on SPC suggests that cultivating a positive mindset can contribute to spiritual fulfillment, a finding that complements existing literature on the interplay between psychological and spiritual well-being63. Furthermore, our study reveals that enhanced PSC facilitates SOC by improving an individual’s social interaction and satisfaction. This finding resonates with Nagy and Carr5 research, which highlights the importance of psychological comforting degree in social contexts. In the tourism industry, for instance, comfort on whale watching vessels significantly impacted tourist satisfaction, with PSC being a crucial factor35. In national parks and mountainous destinations, PHC is influenced by natural elements such as climate, terrain, and available amenities, while PSC is shaped by factors like scenic beauty, solitude and personal safety. The interplay among the four dimensions is vital for creating a holistic comforting experience. For instance, a well-maintained hiking trail not only enhances PHC by offering ease of access but also strengthens PSC by ensuring a sense of security. Similarly, engaging in nature-based activities, such as birdwatching or meditation, fosters SPC by promoting mindfulness and a deeper connection to the environment. Social interactions in these destinations, like group hikes or guided tours, further enhance SOC, underscoring the significance of psychological well-being in group dynamics and communal experiences. Therefore, our results showed that this can be extended to national parks and mountainous destinations, emphasizing the necessity to prioritize PSC to optimize visitor experiences and promote sustainable tourism practices.

Fig. 6
figure 6

Affects the path of recreation comforting experience. *The larger the arrow, the stronger the effect.

Full size image

In this study, PHC exhibited the highest causality (r-c value), as evidenced by the strong directional relationships in (Fig. 4), where arrows consistently point from PHC to other factors, highlighting its crucial role in enhancing overall satisfaction. Pereira et al.25 found that access to fresh air, comfortable temperatures, and other physiological comforts significantly enhance tourists’ psychological and emotional states, thereby strengthening PSC53. Additionally, such physiological comfort experiences contribute to spiritual relaxation and mental purification, reinforcing SPC28. Moreover, Davari64 demonstrated that tourists who experience physiological comfort in natural environments develop a deeper understanding and appreciation of ecosystems, further enhancing their sense of spiritual fulfillment. Furthermore, Zhou et al.26 found that tourists are more likely to engage in positive social interactions in physically comfortable settings, thus improving SOC. These findings collectively support the idea that PHC serves as a fundamental driver of PSC, SPC, and SOC, reinforcing its central role in shaping overall comfort experiences.

Interconnections among sub-criteria of recreation comforting degree

In this study, we identified PHC1 as the pivotal factor in improving overall PHC. This finding resonates with, yet extends, the existing literature that emphasizes the multifaceted benefits of natural environments for human health and well-being65. While previous studies have primarily focused on the indirect effects of natural settings such as promoting physical and mental relaxation65, our research underscores the direct and significant role of PHC1, suggesting it as a primary driver in enhancing tourists’ physical health experiences. Moreover, our results emphasized the direct impact of PHC2. This aligns with and further substantiates existing research on the physiological benefits of outdoor activities and fresh air (e.g., improved immune response and physical fitness). By demonstrating these direct impacts, empirical support has been provided for incorporating specific health promotion activities in national parks and other natural environments. Managers prioritize improving the comfort facilities of the natural environment, such as creating more shaded areas, improving air quality, and developing peaceful walking paths, significantly enhancing the physical health and comfort experience of tourists. In addition, the indirect impact of PHC4 emphasized the importance of incorporating healthy activities into park products. For example, activities such as guided nature walks, outdoor yoga classes, or fitness gymnastics can be introduced. This enhanced their potential as effective health intervention environments.

The pivotal role of PSC1 in enhancing overall PSC was evident from its positive r-c value and its influence on PSC2, PSC3, and PSC4 (Fig. 5). Exposure to and perception of aesthetic elements in both natural and artificial ecological settings likely triggered a sense of harmony and beauty, leading to pleasure and satisfaction. For instance, the visual and olfactory stimuli of tea gardens could stimulate dopamine release in the brain, thereby bolstering an individual’s positive emotional state66. This emotional enhancement is fundamental to psychological comforting degree. Ecological environments, abundant in natural landscapes, sounds, and scents, may have activated deep memory networks, fostering associative memories that intensified emotional experiences and strengthened emotional bonds with the surroundings67. Furthermore, ecological aesthetic experiences have been recognized as vital in adjuvant therapy for psychological disorders. Individuals with depression, characterized by mood swings and anhedonia, may find solace in natural settings. The sensory engagement with natural beauty can evoke positive emotions and memories, potentially mitigating depressive symptoms. The importance of PSC in tourism and hospitality is further underscored by research such as Viñals et al.3which emphasizes the significance of psychological comforting degree in shaping tourist experiences and satisfaction. To create a psychologically comfortable recreational experience, park managers can prioritize the following intervention measures: (1) Developing scenic spots and rest areas to provide stunning natural landscapes. (2) Creating sensory gardens or natural trails through aromatic plants or natural soundscapes, providing a comfortable experience with multiple senses. (3) Provideing commentary and guidance services to highlight the ecological and cultural beauty of the park, thereby cultivating deeper emotional connections and positive psychological experiences.

The enhancement of SPC hinged on SPC1. Ecological knowledge was pivotal for a deeper understanding of the benefits of recreational ecosystem services to humans, facilitating the acquisition of these benefits and enhancing recreational comfort experience64. Park managers can provide educational interpretation programs that increase visitors’ ecological knowledge and deepen their appreciation of the complexity of natural environments. This understanding and appreciation of nature can inspire visitors’ environmental awareness and prompt them to reflect on their lifestyles and values. Creating spaces for reflection and meditation, such as tranquil courtyards or temple pavilions, can further aid in the purification and elevation of individuals’ inner worlds. Additionally, organizing cultural activities that allow visitors to explore local traditions and lives can enhance their appreciation of cultural diversity and promote cultural enrichment. By implementing these initiatives, park managers can effectively enhance SPC, thereby enriching the overall visitor experience.

SOC1, occupying a central role in SOC, underscored the significance of online sharing behaviors in contemporary social interactions. The positive r-c values observed for SOC2 and SOC3 indicated their pivotal roles as driving forces in the establishment of social comforting degree. These factors primarily contributed by bolstering a sense of belonging, offering service support, and facilitating resource sharing, which indirectly reinforced SOC1’s online sharing behaviors. This indirect influence likely operated through the enhancement of individuals’ social connections and trust, thereby stimulating increased online sharing and information exchange55. Despite SOC4’s r-c value being near zero, its contribution to social comforting degree should not be dismissed. SOC4 facilitated individuals in establishing connections with others and sharing travel experiences, indirectly influencing SOC1 by augmenting individuals’ social capital and network resources, ultimately promoting online sharing behaviors. To improve the destination image and social experience of visitors, park managers should focus on social interaction and service quality. First, create an open and friendly social platform to encourage visitors to share their experiences online through activities like photo contests and storytelling events. Second, work with local communities to provide authentic cultural experiences through participation in vibrant festivals and unique events, enhancing visitors’ sense of belonging and social connections. Third, enhance staff training to deliver high-quality services. Fourth, manage visitor numbers dynamically to ensure comfortable personal spaces. To promote spontaneous communication among tourists, designated social interaction areas such as camping area and open-air theaters can be designated.

A data-driven assessment of recreation comforting degree by combining TM with fuzzy DANP

This study introduced a novel approach employing text mining and content analysis to extract key factors influencing recreation comforting degree, contrasting with the literature-derived evaluation dimensions used in most prior studies. This methodology ensures greater objectivity, mitigating the subjectivity and uncertainty inherent in earlier research. By comparing the priority rankings obtained through the DANP method with weights derived from OTA visitor feedback, our findings revealed a consensus on the main criteria ranking (PSC, SOC, SPC, PHC), albeit with discrepancies in sub-criteria rankings. The DANP hierarchy highlighted the significance of emotional and psychological dimensions and the role of social media (e.g., PSC2, SOC1), while OTA feedback emphasized tangible aspects such as natural beauty and service quality (e.g., PSC1, SOC3). To bridge the gap between managerial decisions and tourist experiences, we applied the TM-Fuzzy DEMATEL method, constructing a holistic comfort experience model. The alignment between expert decisions and tourist feedback offers a fresh perspective and decision-support tool for understanding tourist comfort, providing a scientific foundation for national park managers to enhance experiences and services. However, differences in term frequency weights from tourist feedback and expert assignments can be attributed to professional versus experiential perspectives, emotional depth versus pragmatic value, cultural significance, social dynamics, and information asymmetry. The juxtaposition of DANP results and OTA feedback underscores the multifaceted nature of evaluating travel experiences, urging travel planners and policymakers to integrate these diverse views. By doing so, the industry can cater to a wide range of visitor needs, encompassing aesthetic, cultural, physical, and social dimensions.

Conclusion

This study employed the TM-Fuzzy DEMATEL method to bridge the gap between the decision-making of tourism managers and the actual experiences of tourists. It constructed a new decision support model for the holistic comfort experience, encompassing four dimensions: physiological, psychological, spiritual, and social comforting degree. The integration of Fuzzy DEMATEL and DANP methods has unveiled a complex web of interdependencies among the evaluation criteria, underscoring the multidimensionality inherent in tourist comfort. Notably, psychological comforting degree emerged as the central node within this network, exhibiting the highest connectivity and exerting a profound influence on the other criteria. This centrality highlights the pivotal role that psychological comforting degree plays in shaping the overall comfort experience of tourists in national parks. Furthermore, the study’s methodological contribution lies in its integration of both qualitative and quantitative research methods. This dual approach has equipped tourism operators with deeper insights into the needs and perspectives of tourists, enabling a more comprehensive analysis of the diverse factors that influence their comfort experience. The findings offer a scientific basis for national park managers to optimize the tourist experience and improve service quality, ultimately leading to a more sustainable and enjoyable transformational tourism experience for all visitors.

This study has made significant contributions by exploring new methods and applications in the field of multi-criteria decision-making, particularly through the use of text mining technology to analyze big data feedback from tourists on OTAs. By breaking away from traditional methods like Delphi and constructing a novel holistic comfort experience decision support model, the study offers innovative concepts that have the potential to produce breakthroughs in decision-making management. Moreover, the recreation comforting degree, which considers four dimensions (physiological, psychological, spiritual, and social), truly reflects the experiences and satisfaction of tourists, providing valuable insights into their needs and preferences.

In order to further enhance the recreation comforting degree within national parks, the following more targeted strategies are proposed based on the findings of this study. First, the addition of amenities or the development of recreational activities, as well as the creation of a comfortable and harmonious atmosphere, can influence visitors’ comfort experiences. For example, providing comfortable seating and shading facilities can ensure that visitors are able to rest at any time during extended tours, thereby reducing fatigue. National parks can also introduce characteristic recreational activities based on local features, such as tea-picking and ethnic activities, which not only bring a sense of novelty to visitors but also enhance their comfort experiences. In addition, quiet meditation areas can be established within national parks, with simple meditation guidance provided to facilitate visitors’ self-exploration and spiritual relaxation. Secondly, optimizing the services within national parks is also a vital component in enhancing the recreational comfort experience. Employing professional personnel to deliver more vivid and engaging interpretations, and strengthening the training of staff by the park management departments to enhance their service consciousness and professional qualities, can enable them to promptly respond to visitors’ needs and offer warm and considerate services. Additionally, organizing community participation activities can facilitate interactions among visitors and between visitors and local residents, thereby strengthening visitors’ sense of social belonging. Lastly, the integration of digital technology into tourism services is of great significance. Developing a dedicated mobile application allows visitors to access real-time park information, such as introductions to attractions, event schedules, and reservations for dining and accommodation. Simultaneously, visitors can provide feedback on their experiences and suggestions through this application, enabling park management departments to adjust service content and methods in a timely manner based on visitor feedback, thereby further enhancing visitor comforting degree and satisfaction. The introduction of digital technology not only enhances the sense of novelty for visitors but also improves the efficiency and quality of services, making it more convenient for visitors to obtain information and provide feedback, and thus better meeting the needs of visitors.

However, the study also acknowledges certain limitations. The data used in the analysis was sourced solely from OTA reviews, which may not fully represent the experiences of all tourists, particularly those who do not use such platforms. While this study has made important advancements in the field of tourism research by highlighting the significance of psychological comforting degree and developing a new decision support model, there is still room for improvement. Additionally, while the study focused on the role of nature-based national parks in health and well-being of tourists, the results may not be generalizable to other types of tourism experiences or destinations. Future research should aim to address these limitations by exploring a wider range of data sources and considering additional factors that may impact the overall comfort experience of tourists in national parks. Efforts should be made to refine and enhance expert decision-making methods to better align with the real experiences of tourists.