<![CDATA[In the digital era, product reviews on e-commerce platforms such as Amazon have become an important source of information for consumers and sellers. This study develops a system for sentiment analysis of reviews using the Random Forest algorithm and relevant information retrieval with a TF-IDF-based probabilistic model. The data used includes 568,454 product reviews from Amazon, which are processed through data cleaning, tokenization, lemmatization, and feature extraction stages. Sentiments are classified into positive, negative, and neutral. The Random Forest model shows reliable performance with precision, recall, and F1-score of 0.878. The probabilistic search system successfully sorts relevant reviews with a high level of accuracy, which is evaluated using the Mean Average Precision (MAP) metric of 0.878. The results of this study provide significant contributions to improving the e-commerce user experience and supporting data-driven decision making. The approach used opens up opportunities for further research in the fields of natural language processing and machine learning.]]>
Copyrights © 2025
