Table 8 Summary of feature selection and classification algorithm combinations.
Combination | Feature selection | Classification algorithm | Characteristics | Insights |
|---|---|---|---|---|
Combination 1 | Chi square | DT | Categorical attributes, strong associations, robust ensemble. | Chi-Square effectively identifies strong associations, allowing DT to capture non-linear patterns. |
Combination 2 | Information gain | DT | High predictive attributes, ensemble approach. | Information Gain guides DT to leverage informative attributes, potentially improving ensemble performance. |
Combination 3 | Correlation heat map | DT | Linear relationships, interpretability. | Correlation analysis helps DT to handle linear relationships and provides insights into attribute influence. |
Combination 4 | Chi square | RF | Categorical attributes, strong associations, robust ensemble. | Chi-Square highlights attributes with strong associations, allowing RF to capture non-linear patterns. |
Combination 5 | Information gain | RF | High predictive attributes, ensemble approach. | Information Gain guides RF to leverage informative attributes, potentially leading to improved ensemble performance. |
Combination 6 | Correlation heat map | RF | Linear relationships, robust ensemble. | Correlation analysis assists RF in handling linear relationships, enhancing its predictive performance. |
Combination 7 | Chi square | SVM | Categorical attributes, strong associations, non-linear decision boundaries. | Chi-Square identifies strong associations, aiding SVM in addressing complex non-linearity. |
Combination 8 | Information gain | SVM | High predictive attributes, non-linear decision boundaries. | Information Gain guides SVM to focus on informative attributes, potentially improving its ability to handle non-linearity. |
Combination 9 | Correlation heat map | SVM | Linear relationships, non-linear decision boundaries. | Correlation analysis aids SVM in identifying linear relationships, which can be valuable for non-linear decision boundaries. |
Combination 10 | Chi square | NN | Categorical attributes, strong associations, high-dimensional spaces. | Chi-Square highlights attributes with strong associations, allowing NN to work well in high-dimensional spaces. |
Combination 11 | Information gain | NN | High predictive attributes, high-dimensional spaces. | Information Gain guides NN to focus on informative attributes, enhancing its performance in high-dimensional data |
Combination 12 | Correlation heat map | NN | Linear relationships, high-dimensional spaces. | Correlation analysis assists NN in addressing linear relationships and performing well in high-dimensional spaces. |
Combination 13 | Chi square | NB | Categorical attributes, strong associations. | Chi-Square identifies attributes with strong associations, aiding NB’s performance. |
Combination 14 | Information gain | NB | High predictive attributes. | Information Gain helps NB focus on informative attributes, enhancing its performance. |
Combination 15 | Correlation heat map | NB | Linear relationships. | Correlation analysis can aid NB by identifying linear relationships between attributes. |
Combination 16 | Chi square | J48 | Categorical attributes, strong associations, interpretability. | Chi-Square identifies attributes with strong associations, allowing J48 to provide interpretable models. |
Combination 17 | Information gain | J48 | High predictive attributes, interpretability. | Information Gain guides J48 to focus on informative attributes, contributing to interpretable models. |
Combination 18 | Correlation heat map | J48 | Linear relationships, interpretability. | Correlation analysis helps J48 detect linear relationships, enhancing its interpretability. |
