When users engage in online shopping, they often rely on product reviews as a reference. However, efficiently determining the overall evaluation of each product from a large number of reviews is not easy. Previous studies have addressed this issue by automatically scoring each product based on text analysis of product reviews. On the other hand, evaluating the performance of the retail store itself poses challenges. There are usually fewer reviews specifically targeting the store, and evaluations of the store are often embedded within product reviews. As a result, automatically scoring the performance of retail stores is not a straightforward task. Therefore, in this study, we propose a method to extract evaluations of retail stores from product reviews and automatically score them based on store-specific criteria. Additionally, we developed a system that utilizes the calculated store scores, allowing users engaged in online shopping to search for the evaluations of the retail stores that sell the products they are interested in. We present the details of our proposed method and the development of the search system, along with the results of the evaluation experiments conducted using the developed system.

Web advertising services have exhibited consistent growth over the years. However, the conventional methods of web advertising recommendations, relying on keyword matching with search queries and browsing histories, encounter challenges when it comes to effectively targeting users with hidden or latent interests. In contrast, the use of mobile device location data in advertising recommendations often centers around physical store proximity. To address these limitations, our research aims to enhance web advertising recommendations by analyzing latent user interests through real-world behavioral data. This study specifically investigates the influence of area size on user behavioral analysis and its subsequent impact on the accuracy of predicting visit probabilities. We achieve this by extracting the user’s activity range from user behavior (movement) log data and geotagged tweets. Subsequently, we tally the places visited by the user, considering spot attributes, and convert this data into feature vectors. Utilizing these feature vectors in conjunction with various classification methods, we build learning models. In this paper, we present and evaluate these learning models employing different area sizes, verifying their accuracy in predicting user visits to specific stores.

In recent years, running has become increasingly popular as an effective exercise activity which could help improve and maintain one’s physical health. However, even if we understand that exercising regularly has many benefits to our body and brain, it is not easy for people to stay motivated. To deal with this issue, in our previous research, we developed a running support system based on the users’ own records. Our system enables users to run and compete against virtual runners in an audio-based augmented reality space. However, in the AR running support system, competing with real users is more immersive than competing against one’s own running records. In this study, we aim to improve the motivation of running users, not only using the users’ own records, we also use the GPS information and time data from other real runners. We believe that by using information from real runners, our virtual running rival generation/recommendation method can improve the enthusiasm of users. Therefore, we propose a rival recommendation approach for our acoustic-based AR running support system considering the athletic ability of users. In addition, as a preliminary analysis, we conduct a questionnaire survey to clarify users’ needs for the future development of our running support system.

In recent years, there are various review sites on the Web. In online shopping, review sites are important because they strongly influence consumers’ purchasing decisions. We focus on cosmetic reviews and consider the skin type and usability of individual users. In order to realize the cosmetic recommendation system, we are working on the development of a review recommendation method by an automatic scoring system using an evaluative expression dictionary for each cosmetic item classification. Since cosmetic items have detailed classifications, it is necessary to build an evaluative expression dictionary for each cosmetic classification in order to perform automatic scoring. Therefore, it is desirable that the evaluative expression dictionary construction method be efficient and semi-automatic. In this paper, we try to improve the evaluative expression dictionary and examine the efficient method for developing evaluative expression dictionary based on Word2Vec for cosmetics.

Advertisement (ad) recommendation services for mobile users are rapidly increasing. The conventional ways of recommending ads are based on the analysis of user's explicit behavior such as search keywords and keyword matching based on browsing history. However, it might not be effective enough for latent buyers. We have been working on a method to analyze the user's latent interest in web browsing history which categorized positive and negative behaviors. However, we think that the latent interest of users appears not only in the web space but also in the physical space. In this paper, we adapt the method of the linked pages to physical space locations using geo-tagged tweets. Based on several evaluations, we discuss the possibility to recommend ads according to the user's current location.

advertisement (ad) recommendation services for mobile users is rapidly increasing. The conventional ways of recommending ads are based on the analysis of users' explicit behavior such as search keywords and keyword matching based on browsing history. However, it might not be effective enough for latent buyers. We have been working on an analysis of the user's latent interest on web browsing history which categorized positive and negative behaviors. In this paper, we adapt the method of the linked pages to real world locations using geo-tagged tweets. By several evaluations, we discuss the possibility to recommend ads according to the user's current location.

Online advertising has become a popular method for companies to market their products and services to potential customers. The methods used by conventional web advertisement systems to decide on which advertisements to display to users in a real-time bidding environment generally do not consider the latent interests of users and as such it is difficult for advertisers to target and acquire new customers with potential interest in the product. Therefore, we proposed the development of a recommender system which could recommend advertisements to users based on their latent interests. In this paper, we outline two experiment studies related to the development of this system. The first study was carried out to examine the effect of using a long and short browsing history acquisition period to train the user model and predict user interests. The results suggested that a longer browsing history acquisition period did not necessarily result in better predictive performance. The second study examined the use of a categorized web browsing history to predict user interest. The results showed the accuracy of the classifiers increased when website categories were used instead of Fully Qualified Domain Names.

In recent years, although cosmetics review-sharing sites have been helpful in decision making by users, it is not easy for users to find reviews that are suitable for them because the quality of skin and taste vary among individuals. We aim to develop a recommender system for cosmetic items and reviews by analyzing cosmetics reviews. In most review sharing sites, reviewers can assign tags to their own reviews. Tags are very useful for users to understand the effects of the items and to filter reviews with specific tags. Thus, we propose a tag recommendation method for a cosmetics review using the results of the automatic scoring method proposed in our previous work. We believe that our proposed method can significantly simplify the task of assigning tags to a review text. Moreover, the results of the experimental evaluation reveal the tendency that "reviewers may select tags if their score for an aspect of a cosmetic item is sufficiently high". Based on this result, we will discuss a threshold to determine whether to recommend tags.

Portal sites supporting online purchases provide commercial items and reviews for them. In the case of purchasing cosmetic items, in particular, reviews have important roles in purchasing decisions, allowing purchasers to avoid becoming annoyed with unsuitable items. Thus, we are trying to develop a recommender system for cosmetic items and analyzing reviews. General recommender systems basically identify similar users based on their preferences against common items. However, owing to the huge number of cosmetic items, it is not easy to use preferences for common items because of the data sparsity problem. Therefore, we propose a method for finding similar users based on their preferences against cosmetic item clusters. Moreover, we evaluate and discuss the proposed method for finding similar users based on experimental evaluations.

Recent years have witnessed a rapid increase in the use of car navigation systems, which provide drivers with directions to their destinations. However, such systems do not always recommend a route that perfectly matches a driver's intent. Even when drivers intentionally change the driving route from the recommended one to another, most car navigation systems keep recommending or lead them back to the original recommended route. Such recommendations may not adequately reflect a driver's intent. We previously proposed a route recommendation method based on the estimation of a driver's intent by comparing the characteristics of a route selected by a driver and a route not selected by the driver but recommended by the car navigation system. In this study, we propose a method that can consider multiple costs and learn a driver's concept of the values for each cost; in brief, it represents an effective method for learning drivers' route selection preferences. In addition, we describe experimental evaluation results of our proposed method for driving route recommendations and learning drivers' route selection preferences.

Homemaker decide what to cook based on the mood they are in, the ingredients they have in their refrigerators, or the ingredients offered in a supermarket. Most of the existing services for searching recipes allow ingredient names or recipe names as search input. We propose a system that allows searching recipes based on the users' mood. To develop the system, we gather words to express a user's mood when making a menu decision and classify them according to their relationship. We determine six aspects of a user's mood. The result of our preliminary experiment and a questionnaire-based survey show that our method describes a user's mood when deciding for a menu and that the system helps in the decision-making. Furthermore, we propose a method for automatically generating recipe metadata, which we plan to add to our system.

Web advertising is watched with interest as an advertising method employed by companies to introduce their products and services. Web advertising includes listing advertisement, which shows advertisements related to a search keyword, and interest-matching advertising, which shows advertisements relevant to a user's search content and browsing history. However, it is difficult to show effective Web advertising to potential purchasers using the technique based on conventional keyword matching. In this paper, we consider a recommender system for Web advertising based on analysis of the user's potential interests. In particular, we focus on a user model with potential interest for a certain website by analyzing browsing history. We introduce a Web advertising recommender system that is based not only based on keyword matching, but also on reported learning results. In addition, we argue the influence of the period for acquisition of the browsing history, which is taken when the users' model is learned.

Identifying important users from social media has recently attracted much attention in information and knowledge management community. Although researchers have focused on users' knowledge levels on certain topics or influence degrees on other users in social networks, previous works have not studied users' prediction ability on future popularity. In this paper, we propose a novel approach to find important bloggers based on their buzzword prediction ability. We conduct a time-series analysis in the blogosphere considering four factors: post earliness, content similarity, entry frequency and buzzword coverage. We perform preparatory work in categorizing a blogger into knowledgeable categories, identifying past buzzwords, analyzing a buzzword's peak time content and growth period, and finally evaluate a blogger's prediction ability on a buzzword and on a category. Experimental results on real-world blog data consisting of 150 million entries from 11 million bloggers demonstrate that the proposed approach can find prophetic bloggers and outperforms others that do not take temporal features into account.

The purpose of this study is to propose Context-Aware Support Vector Machine (C-SVM) for application in a context-dependent recommendation system. It is important to consider users’ contexts in information recommendation as users’ preference change with context. However, currently there are few methods which take into account users’ contexts (e.g. time, place, the situation and so on). Thus, we extend the functionality of a Support Vector Machines (SVM), a popular classifier method used between two classes, by adding axes of context to the feature space in order to consider the users’ context. We then applied the Context-Aware SVM (C-SVM) and the Collaborative Filtering System with Context-Aware SVM (C-SVM-CF) to a recommendation system for restaurants and then examined the effectiveness of each approach.

Conventional Web bookmarks only contain URLs and titles of Web pages that users are interested in. This makes the process of remembering, sharing or ranking such pages difficult. The "context" of users' navigation can be described as collections of browsed pages. Conventional bookmarks do not contain such information. We believe that such context information conveys the users' intention and the importance of bookmarks. We introduce a notion of context-dependent Web bookmarks that reflects users' browsing histories. A context-dependent Web bookmark consists of (1) representative keywords of bookmarked pages and browsed pages, (2) the ranking value of bookmarked pages calculated by its context, as well as the URL and title of the page that the user bookmarked. Context-dependent bookmarks will make it possible for users to remember the situation of the bookmarking process, grasp the degree of significance of the bookmark, and share the bookmark among multiple users. Furthermore, it becomes possible to re-use context-dependent bookmarks as queries, which could be executed for unvisited Web pages. We also describe our Web browser prototype system based on the context-dependent bookmark function, and our experimental results.