Narrow Results

In today’s simulated path, global and customized music recommendation systems are crucial for reducing the deluge of alternatives, increasing person involvement, and growing a stronger bond between track libraries and their respective listeners. With the purpose of suggestion, a better user enjoys and lives aggressively in the virtual tune marketplace, it is vital to fix this predicament. The biggest obstacle is coming up with a solution that can accurately record all the unique aspects of musical preferences, such as style, environment, speed, and cultural attitude. For the cause of addressing those problems, in this paper, the Fuzzy Convolutional Generative Adversarial Network (FC-GAN) is a modern method that generates customized song recommendations with the aid of combining the strength of deep gaining knowledge of with fuzzy logic. The FC-GAN version makes use of convolutional neural networks (CNNs) to get a whole lot of traits out of audio samples, and fuzzy logic makes it better at taking pictures the subjective nature of people’s adjust preferences. With the use of an adverse education system, FC-GAN is capable of closely ensembling user possibilities with its music embeddings. In the track enterprise and beyond, the counseled FC-GAN structure has first-rate promise for diverse packages. It integrates without problems with existing track streaming systems to provide customized hints, which reinforces user engagement, retention, and happiness. In addition, FC-GAN can make it easier to compose, remix, and generate tune, giving manufacturers and artists extra ingenious tools to specific themselves creatively. By making use of melody datasets and user interaction logs from the actual international, each folks show that FC-GAN can offer extra accurate, diverse, and surprising personalized track recommendations. In addition, the builders validate FC-GAN’s practicality by checking out its scalability, computational efficiency, and resilience throughout distinctive person demographics and musical genres.

There is an abundance of materials for use in professional music courses, but it can be 1 difficult for consumers to quickly and efficiently obtain the specific knowledge they require. Additionally, there is a general lack of data collected on online learning, which renders the recommendation effect of music course resources insufficient. In this study, we use technologies connected to the knowledge graph in the field of online education in order to create a system capable of recommending acceptable educational materials for use in professional music classes. In order to construct a recommendation model using multi-task feature learning, knowledge graphs are embedded within tasks, and high-order connections between latent features and entities are constructed across tasks using cross-compression units. It is possible to achieve success by recommending relevant course materials for individual students based on their requirements, interests and present skill levels. In terms of its ability to generate suggestions, the proposed knowledge spectrogram-based teaching resource recommendation system for professional music courses outperforms four baseline models on a number of publicly available datasets. This method has some practical utility in the domain of course resource suggestion, and its training time is less than that of the comparison model.

Chinese Herbal Medicines (CHM) are the most common interventions of traditional Chinese medicine (TCM), typically administered as either single herbs or formulas. Systematic reviews (SRs) are essential references for evaluating the efficacy and safety of CHM treatments accurately and reliably. Unfortunately, the reporting quality of SRs with CHM is not optimal, especially the reporting of CHM interventions and the rationale of why these interventions were selected. To address this problem, a group of TCM clinical experts, methodologists, epidemiologists, and editors has developed a PRISMA extension for CHM interventions (PRISMA-CHM) through a comprehensive process, including registration, literature review, consensus meeting, three-round Delphi survey, and finalization. The PRISMA checklist was extended by introducing the concept of TCM Pattern and the characteristics of CHM interventions. A total of twenty-four items (including sub-items) are included in the checklist, relating to title (1), structured summary (2), rationale (3), objectives (4), eligibility criteria (6), data items (11), synthesis of results (14, 21), additional analyses (16, 23), study characteristics (18), summary of evidence (24), and conclusions (26). Illustrative examples and explanations are also provided. The group hopes that PRISMA-CHM 2020 will improve the reporting quality of SRs of CHM.

A hybrid protocol is proposed which utilizes secure clustering and hybrid soft computing to improve the network lifetime. Ant colony optimization (ACO) and Particle swam optimization (PSO) with crossover operator are used to design a hybrid soft computing-based inter-cluster data aggregation. Initially, cluster heads are selected based upon adaptive threshold function. Recommendation-based signatures are then assigned to every aggregated data. Then, tree-based data aggregation comes in action and collects sensing information directly from cluster heads by utilizing short distance obtained from the hybrid soft computing. The use of compressive sensing reduces the packet size which is going to be transmitted in the sensor network. Extensive analysis shows that the hybrid protocol considerably enhances network lifetime by conserving the energy in more efficient manner than other protocols at present deployed for sensor networks.

With the merge of digital television (DTV) and the exponential growth of broadcasting network, an overwhelmingly amount of information has been made available to a consumer's home. Therefore, how to provide consumers with the right amount of information becomes a challenging problem. In this paper, we propose an electronic programming guide (EPG) recommender based on natural language processing techniques, more specifically, text classification. This recommender has been implemented as a service on a home network that facilitates the personalized browsing and recommendation of TV programs on a portable remote device. Evaluations of our Maximum Entropy text classifier were performed on multiple categories of TV programs, and a near 80% retrieval rate is achieved using a small set of training data.

e-Tourism is a tourist recommendation and planning application to assist users on the organization of a leisure and tourist agenda. First, a recommender system offers the user a list of the city places that are likely of interest to the user. This list takes into account the user demographic classification, the user likes in former trips and the preferences for the current visit. Second, a planning module schedules the list of recommended places according to their temporal characteristics as well as the user restrictions; that is the planning system determines how and when to realize the recommended activities. Having the list of recommended activities organized as an agenda (i.e. an executable plan), is a relevant characteristic that most recommender systems lack.

In e-commerce applications, the magnitude of products and the diversity of venders cause confusion and difficulty for the common consumer to choose the right product from a trustworthy vender. Although people have recognized the importance of feedback and reputation for the trustworthiness of individual venders and products, they still have difficulty when they have to make a shopping decision from a massive number of options. This paper introduces fuzzy logic into rule definition for users' preferences and designs a novel agent-based decision system using fuzzy rules. This system can help users to find the right product recommendation from a trustworthy vender following users' own preferences.

Modern organizations are keen to work towards their customer needs. To achieve this, analyzing their activities and identifying their interest in any entity becomes important. Every user has been identified as the most important factor in point of organization, and they never give up even a single user. Several approaches have been discussed earlier, which use artificial intelligence to mine the users and their interest in the problem. However, the deep learning algorithms are identified as most efficient in identifying the user interest but suffer to achieve higher performance. Towards this issue, an efficient multi-feature semantic similarity-based online social recommendation system has been proposed. The method uses Convolution Neural Network (CNN) to train and predict user interest in any topic. Each layer has been identified as a single interest, and neurons of the layers are initialized with huge data set. The neuron estimates the Multi-Feature Semantic Similarity (MFSS) towards each interest of the user. Finally, the method identifies the single interest for the user by ranking each interest to produce recommendations to the user. The proposed algorithm improves the performance of recommendation generation with less false ratio.

We present an extended version of the Iterated Prisoner’s Dilemma game in which agents with limited memory receive recommendations about the unknown opponents to decide whether to play with. Since agents can receive more than one recommendation about the same opponent, they have to evaluate the recommendations according to their disposition such as optimist, pessimist, or realist. They keep their first hand experience in their memory. Since agents have limited memory, they have to use different forgetting strategies. Our results show that getting recommendations does not always perform better. With the support of recommendation, cooperators can beat defectors. We observe that realist performs the best and optimist the worse.

Providing recommendations based on distributed data has received an increasing amount of attention because it offers several advantages. Online vendors who face problems caused by a limited amount of available data want to offer predictions based on distributed data collaboratively because they can surmount problems such as cold start, limited coverage, and unsatisfactory accuracy through partnerships. It is relatively easy to produce referrals based on distributed data when privacy is not a concern. However, concerns regarding the protection of private data, financial fears due to revealing valuable assets, and legal regulations imposed by various organizations prevent companies from forming collaborations. In this study, we propose to use random projection to protect online vendors' privacy while still providing accurate predictions from distributed data without sacrificing online performance. We utilize random projection to eliminate the aforementioned issues so vendors can work in partnerships. We suggest privacy-preserving schemes to offer recommendations based on vertically or horizontally partitioned data among multiple companies. The recommended methods are analyzed in terms of confidentiality. We also analyze the superfluous loads caused by privacy concerns. Finally, we perform real data-based trials to evaluate the accuracy of the proposed schemes. The results of our analyses show that our methods preserve privacy, cause insignificant overheads, and offer accurate predictions.

The rapid growth of the World Wide Web has resulted in intricate Web sites, demanding enhanced user skills to find the required information and more sophisticated tools that are able to generate apt recommendations. Markov Chains have been widely used to generate next-page recommendations; however, accuracy of such models is limited. Herein, we propose the novel Semantic Variable Length Markov Chain Model (SVLMC) that combines the fields of Web Usage Mining and Semantic Web by enriching the Markov transition probability matrix with rich semantic information extracted from Web pages. We show that the method is able to enhance the prediction accuracy relatively to usage-based higher order Markov models and to semantic higher order Markov models based on ontology of concepts. In addition, the proposed model is able to handle the problem of ambiguous predictions. An extensive experimental evaluation was conducted on two real-world data sets and on one partially generated data set. The results show that the proposed model is able to achieve 15–20% better accuracy than the usage-based Markov model, 8–15% better than the semantic ontology Markov model and 7–12% better than semantic-pruned Selective Markov Model. In summary, the SVLMC is the first work proposing the integration of a rich set of detailed semantic information into higher order Web usage Markov models and experimental results reveal that the inclusion of detailed semantic data enhances the prediction ability of Markov models.

Today, smartphones are being used to manage almost all aspects of our lives, ranging from personal to professional. Different users have different requirements and preferences while selecting a smartphone. There is ‘no one-size fits all’ remedy when it comes to smartphones. Additionally, the availability of a wide variety of smartphones in the market makes it difficult for the user to select the best one. The use of only product ratings to choose the best smartphone is not sufficient because the interpretation of such ratings can be quite vague and ambiguous. In this paper, reviews of products are incorporated into the decision-making process in order to select the best product for a recommendation. The top five different brands of smartphones are considered for a case study. The proposed system, then, analyses the customer reviews of these smartphones from two online platforms, Flipkart and Amazon, using sentiment analysis techniques. Next, it uses a hybrid MCDM approach, where characteristics of AHP and TOPSIS methods are combined to evaluate the best smartphones from a list of five alternatives and recommend the best product. The result shows that brand1 smartphone is considered to be the best smartphone among five smartphones based on four important decision criteria. The result of the proposed system is also validated by manually annotated customer reviews of the smartphone by experts. It shows that recommendation of the best product by the proposed system matches the experts’ ranking. Thus, the proposed system can be a useful decision support tool for the best smartphone recommendation.

The analysis of the opinions and experiences of tourists is a key issue in tourist promotion. More precisely, forecasting whether a tourist will or will not recommend a given destination, based on his/her profile, is of utmost importance in order to optimize management actions. According to this idea, this research proposes the application of cutting-edge machine learning techniques in order to predict tourist recommendation of rural destinations. More precisely, classifiers based on supervised learning (namely Support Vector Machine, Decision Trees, and $k$-Nearest Neighbor) are applied to survey data collected in the province of Burgos (Spain). Available data suffer from a common problem in real-life datasets (data unbalance) as there are very few negative recommendations. In order to address such problem, that penalizes learning, data balancing techniques have been also applied. The satisfactory results validate the proposed application, being a useful tool for tourist managers.

Recommendation and personalisation approaches aim to filter the most interesting resources that may attract users’ personal interests and preferences by analysing their past attitudes and consumption patterns. The visual aspect of content constitutes an important factor that drives consumers’ attitudes and decisions. In this context, this work proposes a framework for users’ attitude prediction based on items’ visual descriptors and details one of its possible applications for movies recommendations. The main idea of our proposal is to model users’ interests and consumption behaviors using the movies’ posters images and extract features based on the visual descriptors of the items that they interact with, in order to better predict their attitudes towards the ones they do not know. The recommendation approach was integrated into a movies recommendation application that visually assists users while searching for relevant movies to watch, by finding similar movie posters based on the visual aspects of the poster image.

We have developed a PubMed article recommendation system, PURE, which is based on content-based filtering. PURE has a web interface by which users can add/delete their preferred articles. Once articles are registered, PURE then performs model-based clustering of the preferred articles and recommends the highly-rated articles by the prediction using the trained model. PURE updates the PubMed articles and reports the recommendation by email on daily-base. This system will be helpful for biologists to reduce the time required for gathering information from PubMed. PURE is downloadable under GPL license, via http://www.bic.kyoto-u.ac.jp/pathway/mami/out/PURE.tar.gz.

Recommendation in APP becomes increasingly popular as mobile Internet develops rapidly nowadays. Different from music recommendation, tariff package recommendation in telecoms operators’ APP is able to use customers’ recorded information for better performance. Consequently, multi-label learning can be applied for recommendation in that case. During the past decade, multi-label learning has a wide application in real work, such as automatic annotation for multimedia contents, web mining, tag recommendation, etc, and many famous frameworks, such as Binary Relevance(BR). It can raise and solve large amounts of multi-label classification problems successfully. In this paper, a new framework is presented which aims at solving the problem of correlated multi-label classification (CMC), which can be applied in tariff package recommendation. The new framework has three layers. On the first layer, all labels are transformed into several new-generated groups using adapted Principal Components Analysis. Each group is scored on the middle layer and the score will be disperse to original labels on the last layer. Furthermore, we analyze the performance of the framework on tariff package data set. The outcome shows that our framework can obtain more profits with a low time complexity compared to traditional methods.

Nowadays, car navigation systems are widely used to provide drivers with directions to their destinations. However, they do not always recommend a route that perfectly matches the driver's intent. Even when drivers intentionally change the driving route from the recommended one to another, most car navigation systems lead them back to the original recommended route. Such recommendations may not adequately reect the driver's intent. We previously proposed a route recommendation method based on estimating the driver's intent by comparing the characteristics of the route selected by the driver and the route not selected by the driver but recommended by the car navigation system. However, the method only considers one kind of traveling cost for each road. In this paper, we therefore propose a method that can consider multiple costs and learn the driver's concept of the values for each cost.