Abstract :

Customer Reviews Have Become An Essential Source Of Information In Modern E-commerce Platforms, Providing Valuable Insights Into Customer Satisfaction, Product Quality, And User Preferences. However, The Rapid Growth Of Online Shopping Has Resulted In A Massive Volume Of Unstructured Textual Data, Making Manual Analysis Inefficient And Limiting The Effectiveness Of Traditional Review Systems. Existing Approaches Often Rely On Simple Statistical Measures Such As Average Ratings And Review Counts, Which Fail To Capture The Deeper Semantic Meaning And Sentiment Expressed In Textual Feedback. To Address These Limitations, This Study Presents A Multi-task Neural Ensemble Framework For Clothing Review Analysis, Recommendation Prediction, And Rating Estimation. The Proposed Framework Integrates Data Preprocessing, Exploratory Data Analysis, And Feature Extraction Using Term Frequency–Inverse Document Frequency (TF-IDF) To Convert Textual Reviews Into Meaningful Numerical Representations. It Employs A Combination Of Machine Learning And Deep Learning Models, Including Restricted Boltzmann Machines (RBM), Logistic Regression (LR), Ridge Regressor (RR), Gradient Boosting (GB), Extreme Gradient Boosting (XGB), And A Multi-Task Neural Network With Extra Trees (MTNNET). These Models Are Further Enhanced Using The Classification And Regression Tree (CART) Approach To Effectively Handle Both Classification And Regression Tasks Within A Unified Framework. The Experimental Results Demonstrate That The Proposed MTNN-ET-CART Model Achieves Superior Performance, With A Classification Accuracy Of 0.9640 And A Regression R² Score Of 1.0000, Indicating High Prediction Accuracy And Reliability. The Framework Successfully Captures Complex Patterns In Customer Feedback And Generates Precise Recommendation And Rating Predictions.

Published:

09-4-2026

Issue:

Vol. 26 No. 4 (2026)

Page Nos:

2076-2085

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite