Improving Session Based Recommendation by Diversity Awareness

Abstract

Recommender systems help users to discover and filter new and interesting products based on their preferences. Session-Based Recommender systems are powerful tools for anonymous e-commerce visitors to understand their behaviours and recommend useful products. Diversity in the recommendations is an important parameter due to increasing the opportunity of recommending new and less similar items that users interacted. Effect of diversity has been investigated in many works for the collaborative filtering-based Recommender systems. However, for session-based Recommender systems, exploring the effect of diversity is still an open area. In this paper, we propose an approach to calculate the diversity level of the items in the session logs and analyse the effect of diversity level on the session-based recommendation. In order to test the impact of diversity awareness, we propose a sequential Item-KNN recommendation model. The final recommendation list is created as a contribution of the interacted items in the session that depends on the diversity level between last interacted item of the session. We conduct several experiments to validate our diversity aware model on a real-world dataset. The results show that diversity awareness in the sessions helps to improve the performance of Recommender system in terms of recall and precision evaluation metrics. Also, the proposed method can be applied to other sequential Recommender system methods, including deep-learning based Recommender systems.

Personalizing Diversity Versus Accuracy in Session-Based Recommender Systems

“…Anelli et al [55] introduced a variant of xQuAD in which they changed the diversity component in the greedy optimization approach of xQuAD to make it time and session aware. Esmeli et al [56] introduced a variant of IKNN using not only the last item but also the session context in score calculations. They also used the dissimilarity of recommendable items based on the whole session items in the final score function.…”

Using Word2Vec Recommendation for Improved Purchase Prediction

“…Nowadays, users' spend more time on exploring different products and comparing products in different ecommerce platforms to find the most advantageous one in terms of price and quality [8]. Many well-known e-commerce platforms record users' activities and use this data to have personalised content by giving recommendations [13], [14], and purchase prediction in the sessions [43], [44].…”

“…SBRS has a growing trending since their success in providing real-time recommendations even to anonymous users. Many SBRS models are designed [17], [23], [18], [13], [14]. Authors [17] designed RNN based recommendation for short user interaction history their results on Yoochose dataset 1 showed their designed method showed the superiority over the Item-Item similarity-based recommendation.…”

Improving Session-Based Recommendation Adopting Linear Regression-Based Re-ranking

“…One of the ways of re-ranking is taking into account the diversity awareness of the recommended items [21]- [23]. [21] designed a diversity aware re-ranking framework for the session-based recommendation, in which the ranking score of recommended items is updated as based on the results of the diversity level between last interacted item of the session and recommended items. They found that when re-ranking is applied, the RS model produced better performance in terms of recall and precision.…”

“…In this work, we applied a re-ranking approach using a linear regression model on the recommendation list for next item prediction. Re-ranking approaches have been applied in different works [21]- [23], [25], [27], [38] where context-awareness, diversity and popularity based re-ranking options have been used. In this work, we applied a linear regression model to re-rank given the candidate recommendation list, and predict an interest level for recommended items based on multiple factors such as users' behaviours in users' previous sessions and recommended items' features.…”