Search

In the process of user-based collaborative filtering algorithm, finding similar users effectively plays a crucial role in obtaining a high recommendation accuracy. The original rating matrix is very sparse, resulting in similarity information loss during similarity calculating and degrading the efficiency of similar users extracting. To tackle this problem, we propose an improved collaborative filtering algorithm based on Latent Semantic and Bayesian (ICFLSB). ICFLSB first utilizes Latent Semantic to extract meaningful features in the original rating matrix. Then, we establish a Bayesian model based on these extracted features to predict items which users have not rated but may be interested. Further, we fill the sparse original rating matrix with these predicted items and find similar users. After that, we adopt the collaborative filtering algorithm to conduct recommendations. Experiments show that ICFLSB proposed in this paper has a better recommendation performance than the traditional collaborative filtering algorithm. In particular, the evaluation results demonstrate that our ICFLSB can achieve a 2.152% higher and 1.152% higher on recommendation accuracy and recall rate respectively when compared to the traditional collaborative filtering algorithm. [ABSTRACT FROM AUTHOR]

Loopholes exist in nearly every regime, which leads to invalidity and instability in the running process of a regime. In this paper, we propose a new framework based on agent method and game theory to find loopholes in a regime, which contains two players with their clear payoffs. Following the principle of "winners stay, losers change", agents make their own choices among strategies, and their choices can affect the payoffs of other players, resulting in a dynamic equilibrium that has obvious features in decision path, which can be easily found in the final figures after simulation. We further exploit typical cases, such as prisoner's dilemma, anti-coordination game, coordination game and harmony, of which the results have been mathematical proved, to illustrate the validity of our method. Finally, we make some regimes in this paper based on the classical case, "Boxed Pigs", and find the loopholes in these regimes. By utilizing the framework proposed in this paper, managers are able to detect possible problems in advance at the time when they make regimes, which helps reduce the loss of costs caused by management and improve team relations. [ABSTRACT FROM AUTHOR]

With the rapid development of big data and cloud computing, recommender systems (RSs) have gained significant attention in recent decades. However, there are still many challenges and drawbacks existed in RSs, such as complex and high-dimensional data, low recommendation accuracy, time-consuming and low-efficiency, which to a large extent restrict its applications. Non-negative Matrix Factorization algorithm (NMF) is a matrix factorization algorithm which finds the positive factorization of a given positive matrix. It can eliminate invalid and redundant features in user-rating matrix (URM), reduce URM's dimension. Exact Euclidean Locality Sensitive Hashing (E²LSH) is an advanced algorithm for solving the approximate or exact Near Neighbor Search in high dimensional spaces. It can cluster similar-interest users (SIUs) of URM efficiently. Therefore, the authors propose an improved recommender system model named NE-UserCF (NMF-E²LSH-UserCF) based on NMF and E²LSH to improve the quality and performance of recommendation. The authors first utilize the NMF to process original URM, get a new-URM without invalid and redundant features. Then use E²LSH to cluster users in new-URM based on their interests and produce the similar-interest-user matrix (SIUM). The authors further process the Top-10 recommendations by adopting the user-based collaborative filtering algorithm (UserCF). Finally evaluate experimental results by analyzing metrics Precision, Recall, Coverage and Popularity. Experiments indicate that NE-UserCF proposed in this paper improves the quality of recommendation and has a good performance. [ABSTRACT FROM AUTHOR]