Multimedia Content Recommendation Research Articles

Multimedia content recommendation algorithm based on behavior and knowledge feature embedding

Multimedia content recommendation algorithm based on behavior and knowledge feature embedding

An Effective Federated Recommendation Framework with Differential Privacy

This paper proposes a novel federated recommendation framework that incorporates differential privacy to safeguard user privacy without compromising on the accuracy of recommendations. Unlike conventional recommendation systems that centralize user data, leading to potential privacy breaches, our framework ensures that user data remain on local devices. It leverages a federated learning approach, where a global model is trained across multiple devices without exchanging raw data. To enhance privacy protection, we integrate a specially designed differential privacy algorithm that adds carefully calibrated noise to the aggregated data updates. This mechanism ensures that the global model cannot be exploited to infer individual user information. We evaluate our framework on two real-world datasets, one from the e-commerce sector and another from the multimedia content recommendation domain. The results exhibit that our framework achieves competitive recommendation accuracy compared to traditional centralized approaches, with minimal loss in precision and recall metrics, while significantly enhancing user privacy. Our work stands as a testament to the feasibility of creating recommendation systems that do not have to choose between privacy and performance, paving the way for more ethical AI applications in sensitive domains.

EEG Based Emotion Recognition: A Tutorial and Review

Emotion recognition technology through analyzing the EEG signal is currently an essential concept in Artificial Intelligence and holds great potential in emotional health care, human-computer interaction, multimedia content recommendation, etc. Though there have been several works devoted to reviewing EEG-based emotion recognition, the content of these reviews needs to be updated. In addition, those works are either fragmented in content or only focus on specific techniques adopted in this area but neglect the holistic perspective of the entire technical routes. Hence, in this paper, we review from the perspective of researchers who try to take the first step on this topic. We review the recent representative works in the EEG-based emotion recognition research and provide a tutorial to guide the researchers to start from the beginning. The scientific basis of EEG-based emotion recognition in the psychological and physiological levels is introduced. Further, we categorize these reviewed works into different technical routes and illustrate the theoretical basis and the research motivation, which will help the readers better understand why those techniques are studied and employed. At last, existing challenges and future investigations are also discussed in this paper, which guides the researchers to decide potential future research directions.

Application of clustering algorithm in social network user scenario prediction

In order to improve the ability of social network user behavior analysis and scenario pattern prediction, optimize social network construction, combine data mining and behavior analysis methods to perform social network user characteristic analysis and user scenario pattern optimization mining, and discover social network user behavior characteristics. Design multimedia content recommendation algorithms in multimedia social networks based on user behavior patterns. The current existing recommendation systems do not know how much the user likes the currently viewed content before the user scores the content or performs other operations, and the user’s preference may change at any time according to the user’s environment and the user’s identity, Usually in multimedia social networks, users have their own grading habits, or users’ ratings may be casual. Cluster-based algorithm, as an application of cluster analysis, based on clustering, the algorithm can predict the next position of the user. Because the algorithm has a “cold start”, it is suitable for new users without trajectories. You can also make predictions. In addition, the algorithm also considers the user’s feedback information, and constructs a scoring system, which can optimize the results of location prediction through iteration. The simulation results show that the accuracy of social network user scenario prediction using this method is higher, the accuracy of feature registration of social network user scenario mode is improved, and the real-time performance of algorithm processing is better.

Multimedia content recommendation in social networks using mood tags and synonyms

The preferences of Web information purchasers are changing. Cost-effectiveness (i.e., an emphasis on performance with respect to price) is becoming less regarded than cost-satisfaction, which emphasizes the purchaser’s psychological satisfaction. A method to improve a user’s cost-satisfaction in recommending multimedia content is to use the mood-inherent in multimedia items. An example of applications using this method is SNS (Social Network Services) based on mood folksonomy. However, such applications encounter problems due to synonyms. This paper suggests a cost-satisfactory method of multimedia content recommendation to solve the problem of synonyms. It utilizes arousal and valence (AV), which express the mood of multimedia content, as its internal tag. A method that defines the relationship between the AV values of multimedia content and the AV values of mood tags is suggested by considering synonyms and the correlations between them. Furthermore, a multimedia content-recommendation method based on the relationship is suggested and tested. The analysis shows that the AV values of multimedia content with a mood tag and its synonyms reside in the area of the mood tag in the Thayer model. The performance of this method exceeds that of a keyword-based recommendation method.

Frequency-based similarity measure for multimedia recommender systems

Personalized recommendation has become a pivotal aspect of online marketing and e-commerce as a means of overcoming the information overload problem. There are several recommendation techniques but collaborative recommendation is the most effective and widely used technique. It relies on either item-based or user-based nearest neighborhood algorithms which utilize some kind of similarity measure to assess the similarity between different users or items for generating the recommendations. In this paper, we present a new similarity measure which is based on rating frequency and compare its performance with the current most commonly used similarity measures. The applicability and use of this similarity measure from the perspective of multimedia content recommendation is presented and discussed.

A personalized multimedia contents recommendation using a psychological model

With the development and diffusion of compact and portable mobile devices, users can use multimedia content such as music and movie on personal mobile devices, anytime and anywhere. However, even with the rapid development of mobile device technology, it is still not easy to search multimedia content or manage large volume of content in a mobile device with limited resources. To resolve these problems, an approach for recommending content on the server-side is one of the popular solutions. However, the recommendation in a server also leads to some problems like the scalability for a lot of users and the management of personal information. Therefore, this paper defines a personal content manager which acts between content providers (server) and mobile devices and proposes a method for recommending multimedia content in the personal content manager. For the recommendation based on user's personal characteristic and preference, this paper adopts and applies the DISC model which is verified in psychology field for classifying user's behavior pattern. The proposed recommendation method also includes an algorithm for reflecting dynamic environmental context. Through the implements and evaluation of a prototype system, this paper shows that the proposed method has acceptable performance for multimedia content recommendation.

Personalized Recommendation over a Customer Network for Ubiquitous Shopping

Personalization services in a ubiquitous computing environment—ubiquitous personalization services computing—are expected to emerge in diverse environments. Ubiquitous personalization must address limited computational power of personal devices and potential privacy issues. Such characteristics require managing and maintaining a client-side recommendation model for ubiquitous personalization. To implement the client-side recommendation model, this paper proposes Buying-net, a customer network in ubiquitous shopping spaces. Buying-net is operated in a community, called the Buying-net space, of devices, customers, and services that cooperate together to achieve common goals. The customers connect to the Buying-net space using their own devices that contain software performing tasks of learning the customers' preferences, searching for similar customers for network formation, and generating recommendation lists of items. Buying-net attempts to improve recommendation accuracy with less computational time by focusing on local relationship of customers and newly obtained information. We experimented with such customer networks in the area of multimedia content recommendation and validated that Buying-net outperformed a typical collaborative-filtering-based recommender system on accuracy as well as computational time. This shows that Buying-net has good potential to be a system for ubiquitous shopping.