Data Exploration and Modeling using Machine-Learning

Data Exploration and Modeling using Machine-Learning

Guide: Dr. V Shiva Nagaraju, Professor, ECE & IARE

V.Rishitha1, B.Neelima2

1V.Rishitha, ECE & IARE

2B.Neelima, ECE & IARE

Abstract - Data exploration and modelling of Amazon product reviews offers valuable insights into consumer opinions and satisfaction levels, enabling businesses to improve products and services. This study employs natural language processing (NLP) techniques to analyze a large dataset of Amazon reviews, focusing on sentiment polarity (positive, negative, or neutral) and opinion trends. Using advanced machine learning models such as Support Vector Machines (SVM), Naïve Bayes, and deep learning algorithms like LSTM (Long Short-Term Memory), the research achieves high accuracy in sentiment classification. Preprocessing steps, including tokenization, stop-word removal, and stemming, ensure the data is optimized for analysis.

The findings reveal key patterns in customer sentiment, highlighting factors that drive satisfaction or dissatisfaction. Products with high ratings consistently exhibit reviews rich in positive sentiments, emphasizing quality, affordability, and customer service. Conversely, negative sentiments are predominantly associated with issues like delayed shipping, poor durability, or misleading product descriptions. The study also demonstrates how sentiment analysis can identify subtle differences in consumer feedback across product categories, helping businesses tailor strategies to specific markets.

Key Words:  data exploration, data modelling, feedback, percentage, templates, journals