In this study, we propose a method to accomplish two tasks: transforming accumulated data from corporate production activities into knowledge, and recommending necessary information in a timely manner according to corporate activity policies and information users. This method extracts documents containing key phrases from large-scale language resources by leveraging stored text and numeric data. It then generates information expressing intent that encourages the information user’s decision making or some kind of action. Specifically, we propose a method that produces knowledge from past information and makes it reusable. By acquiring new information that serves as a signal for situational changes, the method supports interactive communication and recommends information that aids users in making personalized decisions within corporate activities. Specifically, while showing examples of the application of specialized knowledge domains in a specific customer segmentation within corporate activities, this method is centered on language and other numeric information generated within corporate economic activities that are recorded on databases. To express information that either represents the relationship between customers and companies or prompts decision-making or behavioral changes, domain-specific knowledge is built from various knowledge information processing tasks within economic activities. Then, through information acquisition that detects signals based on situational awareness, this method indicates that each text can be generalized into semantic representations as personalized information for users.

While individuals benefit from the goods and services provided by companies that enrich their lives and that have adapted to a dynamic environment that is always changing, these companies pay a high communication cost to access opportunities to provide these goods and services and to seek a better understanding of individual customers’ changing needs. Although vast amounts of information can be obtained, databases and machine learning are playing an increasingly important role in extracting meaning from this information, turning it into meaningful information assets that consider circumstances and contexts, and individualizing the economy of information. I propose an implementation method for providing information to enrich the profiles of individual customers by consolidating different data, calculating the individual customers’ needs through the relationships between customers and products, evaluating the change in relationships between individual customers and products over time, and providing goods and services to suit different intervals of change to factors such as lifestyle and living environment. As there are different factors involved in estimating the incidence of needs, and different frequencies and rates at which they occur, based on the special characteristics of products, different data are required to estimate such needs. By profiling individuals over the long term, it is possible to build an information provision environment that is conducive to companies’ customer acquisition.

Improved computing environments performing large-scale data processing and high-speed computational processing facilitate the delivery of new algorithms to businesses while considering cost efficiency for small-scale investments. Implementing the proposed method more as a criterion for feasibility and economic rationality in specific problem areas rather than as an approach to generic issues, we aim to develop technologies of practical use in the real world. Recently, it has become possible for customers to monitor their buying behavior through smart devices, and with the improvement of computing performance, it has become possible to improve the accuracy of prediction and recommendation cycles through active online learning. This study proposes a method for dynamically recommending products that are highly likely to be selected by the user by combining the user's reaction with reuse of knowledge and real-time online learning to cyclically repeat feedback that is more specific to the user.We propose a method to sense streaming data by utilizing a user's behavior, intervening a user's behavioral change through interactions, such as recommendations, and evaluating the userâĂŹs buying intention and interest in each product. Using the evaluation results for recommendations helps achieve positive feedback and effectively support the selection of more exciting or different products. We propose a recommendation method specific to individual customers based on past transaction data, where changes can be monitored in real-time by reusing the knowledge acquired in advance through batch processing of knowledge discovery and data mining and processing the stream data in real-time online. We will present the implementation of our proposed method targeting the database system and machine learning algorithm.

We herein present a method that dynamically generates the curricula specialized to the learning circumstances of individual learners, given prior learning goals and learning objects. A generated curriculum encourages the selection of learning behaviors according to the learning objects that may be feasibly acquired within a limited timeframe. Our method evaluates the circumstances of an individual learner; by dynamically selecting feasible learning objects based on the individual's learning behaviors and their past records, it finds the best learning tasks within the constraints of time, circumstances, and activities. Using prior known rules and strengths of causal/dependency relations between learning items, our method enables, from individual test results, the discovery of the learning objects that are important, and how they should be ordered, on an individual basis. This enables an effective support in choosing the most appropriate learning behaviors, tailored to the individual learner. It also enables the selection of effective learning behaviors by examining the behavior records of other individuals, treating the influence of their prior learning behaviors on subsequent learning behaviors as experience quotients and using them by converting them into expected scores for the individual's learning behaviors. Accordingly, we evaluate whether the learning behaviors selected by the individual are indeed learning tasks that would correspond to anticipated learning results, conduct prior assessment of the influence that the results of this intervention would have upon the learning circumstances and thus prioritize more effective learning behaviors. When implemented, our method assesses the changes in the individual's learning circumstances based on their learning behaviors on a timeline and subsequently adjusts the recommended behaviors. The method can provide effective support for individual learners: along with effective feedback on learning task selection in response to the individual's circumstances, it dynamically generates an individualized curriculum by measuring the relationships between the individual's learning circumstances and learning items. We herein present a method for dynamically generating the curricula in response to an individual's learning circumstances through measuring the causal/dependency relations between learning items, thus enabling the calculation of the relationship between an individual's past learning record, and the learning behaviors and learning objects available to be chosen by the individual. We investigate its efficacy and achievability through empirical testing by using actual data.