О рекомендательной маршрутной системе, основанной на оценке предпочтений пользователя
The invention relates to a method for selecting and ranking valid variants in search and recommendation systems. The claimed method makes it possible to select and rank variants with a high degree of accuracy and speed, especially in the case of a large number of variants which can be characterized by a large set of indicators. In the claimed method (variant), criteria for evaluating the relevance of a variant to the search request are first generated and a set of procedures for the selection and ranking of variants and a sequence for performing said procedures for the selection of variants evaluated as the most valid are established, and an evaluation of each of the variants is made on the basis of the relevance to search request criteria and, on the basis of this evaluation, the variants are ranked by means of assigning a rank to each of said variants based on the condition of correspondence to the greatest number of criteria in decreasing order, and then the variants are selected and ranked in at least two stages using the superposition method, and the variants are selected, ranked and excluded until all of the established selection procedures have been used and the selected group of variants is evaluated as being the most valid.
Training model information recommendation system is associated with the study of applied mathematical and information methods and models, their combinations in order to ensure the necessary accuracy of the forecasts and conclusions. The article deals machine learning of model recommendation system using statistical methods and analysis of big data, aimed at addressing the issues of individualization of education. In this case, the accuracy of the machine learning model depends on the type of statistical model used to predict the probability of some event from the values of the set of features, as well as the training sample used to select the parameters, and the regularization function used to improve the generalizing ability of the resulting model. The study tested models based on logistic regression, methods of naive Bayesian classifier (Naïve Bayes), lasso-type regression. Experimentally confirmed the theoretical assumption about the possibility of creating a recommendation system on the individualization of education on the basis of an array of educational data, including the results of educational and extracurricular activities of students. Conclusions about the presence of correlation dependencies in the data, which can be used to improve the accuracy of the model of the recommendation system, are formulated. Keywords: individualization of education, machine learning models, big data analysis, recommendation systems
This paper explains how people responding to our survey, which included users’ basic information, social status, experience with social networking and attitude towards social network-integrated e-health information systems. The survey findings show that social media users need special recommendation and guidance services—especially those people located in urban centers that have busy schedules. These people prefer to receive recommendations for their minor health problems over having to go to the hospital or clinic and spend time waiting, perhaps even to return home without a proper consultation from a doctor. As a result, we propose to work on architecture for integrated social media analytics and e-health information systems. However, our findings, being the result of a controlled survey, raise issues such as respondent trust and security and privacy issues relating to healthcare.