Anzeige
Erschienen in:
2022 | OriginalPaper | Buchkapitel
Session Based Recommendations using CNN-LSTM with Fuzzy Time Series
verfasst von : Punam Bedi, Purnima Khurana, Ravish Sharma
Erschienen in: Artificial Intelligence and Speech Technology
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Abstract
Session based Recommender systems consider change in preferences by focusing on user’s short term interests that may change over a period of time. This paper proposes FS-CNN-LSTM-SR, a hybrid technique that uses CNN (Convolutional Neural Networks) and LSTM (Long Short Term Memory) deep learning techniques with fuzzy time series to recommend products to user based on his activities performed in a session. The advantage of our proposed method is that it combines the benefits of both CNN and LSTM. CNNs are capable of extracting complex local features and LSTM learn long term dependencies from sequential session data. The performance of FS-CNN-LSTM-SR is evaluated on YOOCHOOSE dataset from RecSys Challenge 2015 and is compared with three variations viz. LSTM-SR, CNN-LSTM-SR and FS-LSTM-SR. We observed that our proposed approach performed better than other three variations. The proposed technique is applicable on any E-commerce dataset where user purchasing choices need to be predicted.