Table 1.
A comparative analysis of related works on predicting purchase behavior using machine learning.
Fig 1.
Methodology of the study.
Table 2.
Demographic profile of respondents.
Fig 2.
Gender-wise distribution amongst participants.
Fig 3.
Distribution of occupation across different ethnic group.
Table 3.
Statistically significant differences between early and late respondents.
Table 4.
Selected TPB factors and definitions with respective feature set.
Table 5.
Categorical values encoded with corresponding numerical values.
Fig 4.
Responses from attitude factors.
Fig 5.
Responses from social norm factors.
Fig 6.
Responses from perceived behavioral control(PBC) factors.
Fig 7.
Responses from purchase behavior factors.
Fig 8.
Percentage distribution(a) of respondent on purchase behavior(b) by gender.
Fig 9.
Percentage distribution(a) of respondent on purchase behavior(b) by age.
Fig 10.
Percentage distribution(a) of respondent on purchase behavior(b) by attention to the advertisement on social media(ADSM).
Table 6.
Parameter space and best parameters of models.
Table 7.
Accuracy table of purchase behavior (PB) on all possible combinations.
Table 8.
Macro F1 scores on all possible combinations.
Fig 11.
Radar chart for the best model in terms of accuracy and macro F1 score.
Fig 12.
Confusion matrix of the best performing model in each combination of TPB factors.
Fig 13.
ROC curves of the best performing models in each combination of TPB factors.
Fig 14.
Precision recall curves of the best performing models in each combination of TPB factors.
Fig 15.
The interpretation of purchase behavior prediction of low influence case using LIME explainable AI.
Fig 16.
The interpretation of purchase behavior prediction of high influence case using LIME explainable AI.