Probabilistic neural network approach to alleviate sparsity and cold start problems in collaborative recommender systems
32
Citation
10
Reference
10
Related Paper
Citation Trend
Abstract:
Collaborative Recommender system helps the online users to identify the right product during electronic purchasing. The collaborative recommender system identifies the similar users based on the purchasing or rating behavior to the active user and then recommends the product based on the similar users. Collaborative recommender system is widely used in majority of the existing online recommender system such as orkut, google, amazon, walmart etc. Besides it popularity, is suffers due to sparsity, cold start and scalability recommender system. Extensive research is going on to overcome these problems. In this paper, Probabilistic neural network (PNN) is used to calculate the trust between users based on rating matrix. Using the calculated trust, sparse rating matrix is smoothened, by predicting the rating values of the nonrated items in the rating matrix. Using this smoothened rating matrix, the trust is calculated for online active users. The calculated trust is used to recommend product. Experiments are conducted using dataset such as movielens. Based on the performance metrics, it is proved that the proposed method performs better than the benchmark and some existing systems.Keywords:
MovieLens
Benchmark (surveying)
Cold start (automotive)
Collaborative filtering algorithm is one of the most widely used algorithms in recommender systems and has demonstrated promising results. But it relies too much on similarity to find the nearest neighbors. Whatever, the trust between users is also an import factor needed to be considered. This paper proposed a collaborative filtering algorithm that combined the user similarity and trust to obtain a more appropriate nearest neighbors set. Users not only have same interests as their nearest neighbors, but also have higher level of acceptance in the items recom-mended by their nearest neighbors. Extensive experiments based on Film Trust and MovieLens datasets have shown that the approach has major potential in improving the accuracy of recommended item.
MovieLens
Similarity (geometry)
Cite
Citations (7)
Collaborative filtering algorithm is one of the most widely used algorithms in recommender systems and has demonstrated promising results. But it relies too much on similarity to find the nearest neighbors. Whatever, the trust between users is also an import factor needed to be considered. This paper proposed a collaborative filtering algorithm that combined the user similarity and trust to obtain a more appropriate nearest neighbors set. Users not only have same interests as their nearest neighbors, but also have higher level of acceptance in the items recom-mended by their nearest neighbors. Extensive experiments based on Film Trust and MovieLens datasets have shown that the approach has major potential in improving the accuracy of recommended item.
MovieLens
Similarity (geometry)
Cite
Citations (0)
Collaborative filtering is one of the most extensive and successful personalized recommendation algorithm in e-commerce recommendation system.Affected by data sparsity,the traditional collaborative filtering algorithms does not reflect the interest similarity of uses calculating similarity between users on the smaller set of common rated items accurately,seriously affecting the accuracy of recommendation system.To solve this problem,collaborative filtering algorithm based on co-ratings was proposed by analyzing the distribution of co-ratings and relationship between co-ratings and similarity,directly using co-ratings as a criterion to select nearest neighbor without calculating similarity.Experiments on MovieLens datasets show that the algorithm can make a substantial increase in prediction accuracy and recommendation coverage.
MovieLens
Similarity (geometry)
Data set
Cite
Citations (0)
Collaborative filtering (CF) is one of the most popular recommender system technologies. It tries to identify users that have relevant interests and preferences by calculating similarities among user profiles. The idea behind this method is that, it may be of benefit to one's search for information to consult the preferences of other users who share the same or relevant interests and whose opinion can be trusted. However, the applicability of CF is limited due to the sparsity and cold-start problems. The sparsity problem occurs when available data are insufficient for identifying similar users (neighbors) and it is a major issue that limits the quality of recommendations and the applicability of CF in general. Additionally, the cold-start problem occurs when dealing with new users and new or updated items in web environments. Therefore, we propose an efficient iterative prediction technique to convert user-item sparse matrix to dense one and overcome the cold-start problem. Our experiments with MovieLens and book-crossing data sets indicate substantial and consistent improvements in recommendations accuracy compared with item-based collaborative filtering, singular value decomposition (SVD)-based collaborative filtering and semi explicit rating collaborative filtering.
MovieLens
Cold start (automotive)
Cite
Citations (15)
Collaborative filtering (CF) algorithms have received a lot of interest in recommender systems due to their ability to give personalized recommendations by exploiting user-item interaction data. In this article, we explore two popular CF methods—K-Nearest Neighbors (KNN) Regression and Non-Negative Matrix Factorization (NMF)—in detail as we dig into the world of collaborative filtering. Our goal is to evaluate their performance on the MovieLens 1M dataset and offer information about their advantages and disadvantages. A thorough explanation of the significance of recommender systems in contemporary content consumption settings is given at the outset of our examination. We look into Collaborative Filtering's complexities and how it uses user choices to produce tailored recommendations. Then, after setting the scene, we explain the KNN Regression and NMF approaches, going over their guiding principles and how they apply to recommendation systems. We conduct an extensive investigation of KNN Regression and NMF on the MovieLens 1M dataset to provide a thorough evaluation. We describe the model training processes, performance measures, and data pre-processing steps used. We measure and analyse the predicted accuracy of these strategies using empirical studies, revealing light on their effectiveness when applied to various user preferences and content categories.
MovieLens
Non-negative Matrix Factorization
Empirical Research
Cite
Citations (0)
Recommender systems have become significant tools in electronic commerce, proposing effectively those items that best meet the preferences of users. A variety of techniques have been proposed for the recommender systems such as, collaborative filtering and content-based filtering. This study proposes a new hybrid recommender system that focuses on improving the performance under the "new user cold-start" condition where existence of users with no ratings or with only a small number of ratings is probable. In this method, the optimistic exponential type of ordered weighted averaging (OWA) operator is applied to fuse the output of five recommender system strategies. Experiments using MovieLens dataset show the superiority of the proposed hybrid approach in the cold-start conditions.
MovieLens
Cold start (automotive)
Fuse (electrical)
Cite
Citations (34)
The rating and review based algorithms like collaborative filtering faces the cold start problem on addition of new items or new users in the recommender system because of lack of ratings and reviews. The primary aim of this work is to solve the cold start problem faced by rating based recommendation system and also increase the accuracy of recommender systems. We used the content-based filtering algorithm along with the cross-domain data taken from Facebook so that the problem of addition of a new user problem and a new item in recommender system can be solved. The combination of both content-based filtering and cross-domain data obtains better results than the collaborative filtering and shows a good way to deal with the cold start problem.
Cold start (automotive)
Cite
Citations (4)
MovieLens
Cold start (automotive)
Cite
Citations (1)
Collaborative Filtering is the most successfully and widely used recommendation algorithm nowadays. However, it faces severe challenges of the cold-start problem. This paper proposes a hybrid algorithm by using the users' attributes and the back propagation neural network (BPNN) to solve the new user problem in cold-start. Firstly, users' attributes are collected and analyzed to calculate the component similarity respectively. Secondly, take these similarities as the input neurons of the BPNN and set the initial values of weight and biases parameters, through the self-adjusting and self-adaptive features of the BPNN, optimize the rating prediction model to get a higher recommendation accuracy of the algorithm. Finally, take the output as the predicted similarity of the new user, so that the algorithm can recommend for the new user. After a series of experiments carried out on a public data set -- MovieLens(ml-1m), the experimental results show that the proposed method can effectively accomplish novel user recommendation for collaborative filtering cold-start and reduce the performance loss.
MovieLens
Cold start (automotive)
Similarity (geometry)
Component (thermodynamics)
Cite
Citations (15)
협업여과는 추천시스템에서 널리 사용되는 기법으로 다른 사용자의 평가를 기반으로 아이템을 추천하는 기법이다. 사용자 데이터베이스를 이용하는 메모리기반 협업여과에는 사용자기반 기법과 아이템기반 기법이 있다. 사용자기반 협업여과는 유사한 선호도를 가지는 이웃사용자들의 선호도를 바탕으로 특정 아이템에 대한 선호도를 예측하는 반면, 아이템기반 협업여과는 아이템들의 유사도를 바탕으로 특정 사용자의 선호도를 예측한다. 본 논문에서는 추천의 성능을 향상시키기 위하여 이웃사용자와 이웃아이템 크기의 비율을 가중치로 하여 사용자기반 예측값과 아이템기반 예측값을 결합함으로써 최종 예측값을 생성하는 결합예측기법을 제안한다. MovieLens 데이터 셋과 BookCrossing 데이터 셋을 이용한 실험을 통해 본 논문에서 제안한 결합예측기법이 영화와 책에 대하여 사용자기반과 아이템기반보다 예측의 정확성을 향상시킴을 보인다. Collaborative filtering is a popular technique that recommends items based on the opinions of other people in recommender systems. Memory-based collaborative filtering which uses user database can be divided in user-based approaches and item-based approaches. User-based collaborative filtering predicts a user's preference of an item using the preferences of similar neighborhood, while item-based collaborative filtering predicts the preference of an item based on the similarity of items. This paper proposes a combined forecast scheme that predicts the preference of a user to an item by combining user-based prediction and item-based prediction using the ratio of the number of similar users and the number of similar items. Experimental results using MovieLens data set and the BookCrossing data set show that the proposed scheme improves the accuracy of prediction for movies and books compared with the user-based scheme and item-based scheme.
MovieLens
Similarity (geometry)
Data set
Cite
Citations (0)