Table 8 Summary of multiple linear regression models predicting overall perception in Suzhou (N = 20) and Kyoto (N = 20)

From: Toward sustainable and differentiated protection of cultural heritage illustrated by a multisensory analysis of Suzhou and Kyoto using deep learning

  

Suzhou

Kyoto

  

Estimate

Estimate

Text-Natural element

Greening

0.3466

0.1062

 

Flower

0.0673

0.1043

 

Water/Mountain/Stone

−0.032

0.0779

 

Animal

−0.0267

0.1388

 

Natural phenomenon

−0.0415

−0.0184

 

Season

0.0755

0.0942

  

Pseudo R²:0.22

Pseudo R²:0.36

Artificial element

Architecture

0.1366

0.1162

 

Cultural facilities

−0.2453

−0.0115

 

Landscape structure

0.0104

0.1288

 

Space/interior

−0.055

0.0271

 

Service

−0.2424

−0.0886

 

Cost/Fee

−0.0898

−0.0812

 

FestivaI Activities

0.1221

0.0622

 

Crowd

−0.0298

0.1027

  

Pseudo R²: 0.3

Pseudo R²: 0.27

Multi-sensory

Vision

0.4069

0.107

 

Hearing

−0.0635

0.1571

 

Olfactory/taste

−0.1286

0.0448

 

Feeling

−0.0875

0.0651

  

Pseudo R²: 0.29

Pseudo R²: 0.253

Photos

Sky

0.0053

0.1123

 

Vegetation

−0.02

0.1031

 

Mountain

0.0278

−0.1747

 

Water

0.0583

−0.2333

 

Architecture

−0.0211

−0.0993

 

Interior

−0.0292

−0.0939

 

Structure

0.1164

−0.3388

 

Road

−0.0785

−0.173

 

Transportation

0.4476

−1.1192

 

People

−0.0516

−0.2333

  

Pseudo R²: 0.214

Pseudo R²: 0.253

  1. Note. Traditional R² reflects the proportion of variance explained in linear models, while Pseudo R² is used in generalized linear models to indicate relative model fit. Pseudo R² values above 0.2 suggest acceptable explanatory power.
Back to article page