An Adaptive Model for Catalyzing Digital Marketing Using Machine Learning

نوع المستند : المقالة الأصلية

المؤلفون

1 جامعة حلوان

2 كلية الحاسبات والذكاء الاصطناعى جامعة حلوان

3 كلية التجارة وادارة الاعمال-جامعة حلوان-قسم ادارة الاعمال

4 نظم المعلومات - كلية التجارة وإدارة الأعمال - جامعة حلوان.

المستخلص

In today’s digital era, the vast amount of unstructured data generated daily from different data sources like social media, e-commerce platforms, emails, blogs, and other online sources has made sentiment analysis a critical tool for businesses. This paper aims to analyze Amazon customer reviews using a three-tier approach: (1) data preprocessing, (2) statistical feature selection to identify key variables, and (3) classification with machine learning algorithms—Random Forest, Naïve Bayes, and SVM. The classifiers are evaluated using performance metrics such as accuracy, precision, F-measure, true positive rate, and false negative rate. A comparative analysis reveals their strengths and limitations, providing actionable insights for optimizing sentiment analysis. This paper enhances sentiment analysis through structured processing of unstructured reviews, providing businesses with actionable insights involving stacking ensemble combining Naïve Bayes, Gradient Boosting, Neural Networks under an XGBoost for sentiment analysis to drive data-based decisions and improve customer satisfaction.

الكلمات الرئيسية

ملفات

شارك

إرسال الاستشهاد إلى

الإحصائيات