Narrow Results

The clustering coefficient of user–object bipartite networks is presented to evaluate the overlap percentage of neighbors rating lists, which could be used to measure interest correlations among neighbor sets. The collaborative filtering (CF) information filtering algorithm evaluates a given user's interests in terms of his/her friends' opinions, which has become one of the most successful technologies for recommender systems. In this paper, different from the object clustering coefficient, users' clustering coefficients of user–object bipartite networks are introduced to improve the user similarity measurement. Numerical results for MovieLens and Netflix data sets show that users' clustering effects could enhance the algorithm performance. For MovieLens data set, the algorithmic accuracy, measured by the average ranking score, can be improved by 12.0% and the diversity could be improved by 18.2% and reach 0.649 when the recommendation list equals to 50. For Netflix data set, the accuracy could be improved by 14.5% at the optimal case and the popularity could be reduced by 13.4% comparing with the standard CF algorithm. Finally, we investigate the sparsity effect on the performance. This work indicates the user clustering coefficients is an effective factor to measure the user similarity, meanwhile statistical properties of user–object bipartite networks should be investigated to estimate users' tastes.

In this paper, we empirically investigate the statistical properties of the user correlation network in terms of their common rated objects on MovieLens, and find that it has high clustering coefficient and ultra small average distance, which is close to the fully connected network. We argue that the above characteristics come from the fact that large-degree objects build lots of fully connected subnetworks by using the node projection method. By introducing the user global similarity, measured by the product of two users' similarity vectors, we present an effective way to identify users' specific interests by weakening the mainstream interests and noise interests. Numerical results show that we are able to obtain accurate and diverse recommendations by considering the second-order correlation redundant information simultaneously, which outperforms the state-of-the-art collaborative filtering (CF) methods. This work suggests that statistical properties of the user correlation network is an important factor to improve the performances of information filtering algorithms.

The process of heat conduction (HC) has recently found application in the information filtering [Zhang et al., Phys. Rev. Lett.99, 154301 (2007)], which is of high diversity but low accuracy. The classical HC model predicts users' potential interested objects based on their interesting objects regardless to the negative opinions. In terms of the users' rating scores, we present an improved user-based HC (UHC) information model by taking into account users' positive and negative opinions. Firstly, the objects rated by users are divided into positive and negative categories, then the predicted interesting and dislike object lists are generated by the UHC model. Finally, the recommendation lists are constructed by filtering out the dislike objects from the interesting lists. By implementing the new model based on nine similarity measures, the experimental results for MovieLens and Netflix datasets show that the new model considering negative opinions could greatly enhance the accuracy, measured by the average ranking score, from 0.049 to 0.036 for Netflix and from 0.1025 to 0.0570 for Movielens dataset, reduced by 26.53% and 44.39%, respectively. Since users prefer to give positive ratings rather than negative ones, the negative opinions contain much more information than the positive ones, the negative opinions, therefore, are very important for understanding users' online collective behaviors and improving the performance of HC model.

The Internet is a vast, semi-structured information repository that is expanding ever-faster with two alarming results. First, finding information is becoming increasingly difficult, and second, the management of such information resources is also becoming more complex. We have developed an Intelligent Distributed Information Management System (IDIoMS) to solve these problems. IDIoMS provides a comprehensive set of tools for personalized information delivery, and a platform that enables straightforward management of distributed information resources using agent technologies. These agents enable the sharing, management, search, and presentation of useful information that is widely distributed over a network such as the Internet. IDIoMS benefits users through the automatic provision of timely and relevant information with minimal need for users to search for that information. In addition, the system also benefits service providers through the plug-and-play provision of information services that minimizes the overhead needed to make their service widely available. This paper describes the concept and the architecture of IDIoMS and discusses its utility through three field trials.

In this paper, we formulate the recommendation problem as a hybrid combination of one-class classification with collaborative filtering. Specifically, we decompose the recommendation problem into a two-level cascade scheme. In the first level, only desirable items are selected for each user from the large amount of all possible items, taking into account only a small portion of his/her available preferences. This is achieved via a one-class classification scheme trained only with positives examples, i.e. only with desirable items for which users have provided a rating value. In the second level, a collaborative filtering approach is applied to assign a rating degree to the items identified at the first level. The efficiency of our approach is analyzed theoretically in terms of best/worst case scenarios and respective lower/upper mean absolute error (MAE) bounds are computed. Moreover, our approach is experimentally tested against pure collaborative and cascade content-based approaches. The results show that our approach outperforms them in terms of MAE and, moreover, the experimental MAE is close to the theoretical lower bound corresponding to the best case scenario. The superiority of our approach is due to the existence of the one class classifier in the first level of the cascade.

Recommender Systems (RS), allow users to share information about items they like or dislike and obtain, in a timely fashion, recommendations based on predictions about unseen items (physical or information goods and/or services). In this process, users' preferences are considered to be the learning target functions. We study Agent-based Recommender Systems (ARS) under the scope of online learning in Multi-Agent systems (MAS). This approach models the problem as a pool of independent cooperative predictor agents, one per each user (the masters) in the system, in situations in which each agent (the learners) faces a sequence of trials, with a prediction to make in every step, eventually getting the correct value from its master. Each learner is willing to discover the degree of similarity among the target function of its master and those of other agents' masters (i.e. preference similarity). The agent uses this information for the calculation of its own prediction task, the goal being to make as few mistakes as possible. A simple, yet effective method is introduced in order to construct a compound algorithm for each agent by combining memory-based individual prediction and online weighted-majority voting. We give a theoretical mistake bound for this algorithm that is closely related to the total loss of the best predictor agent in the pool. Finally, we conduct some experiments obtaining results that empirically support these ideas and theories.

This paper develops a Natural Language Processing method for extracting temporal information of events from news articles. The extraction process is based on the result of a morpho-syntactic analysis. The obtained results, which are a translation of morpho-syntactic sentences, are skimmed in order to identify temporal markers.

In our approach, we consider a specific description of events. Our approach is based on a methodology of temporal markers classes and on a contextual exploration method.

Our experimental results indicate that our approach is significant for extracting relevant event information. We have also achieved a system, called EXEV, on the basis of the proposed approach and have empirically verified its effectiveness. More specifically, EXEV system automatically extracts all information about events (paragraphs or sentences) from news articles and specifies more information about events: associations, locations, temporal settings, etc.

The system was tested on press articles and gives recall and precision more than 80%.

Due to the explosive growth of available information on the World Wide Web (WWW), users have suffered from the information overload. To alleviate this problem, there is a need for an intelligent tool to help the users screening and filtering for interesting and useful information. In this paper, a method of automatically identifying topics for Web documents via a classification technique is proposed. Topic identification can be applied as a filtering tool for recommender systems to prune down the number of documents to within some particular topics. We adopt the fuzzy association concept as a machine learning technique to classify the documents into some predefined categories or topics. Our approach is compared to the vector space model with the cosine coefficient using the data sets collected from three different Web portals: Yahoo!, Open Directory Project and Excite. The results show that our approach yields higher classification accuracy compared to the vector space model.