DeNet_SVM: Product Based Recommendation System using Deep Learning and Web Usage Mining
Bhavana Bhavana, Neeraj Raheja · 2022
The world is abundant with digital products in which selecting, searching and buying products for the users is harder. Recommendation system helps the users to overcome this drawback and improves the users’ experience. Recommendation systems increase the outcomes of the companies by providing a list of favourite products to the users. This paper focuses on providing efficient recommendation to the users according to the session constructed for every user based on web usage. This paper proposed product-based recommendation system on image dataset using deep learning and web usage mining. In this paper, MobileNetV2, DenseNet201 are used as feature extraction and SVM, NB, Ensemble, KNN are used as classifiers. DeNet_SVM achieved 92.59% accuracy and 77.76% F-Measure. DenseNet201 and SVM i.e., DeNet_SVM gives better results in terms of accuracy, Precision, Recall and F-Measure in comparison to other techniques. The experiment is performed on standard dataset myntra.dataset which contains 600 images classified into 6 different categories.