Intelligent search for scientific information in e-learning systems

Abstract

The advantages of using machine learning in search are that the search engine can learn and thus lead to more personalized answers, rather than the most popular results. In well-known search engines, such algorithms have been used for a long time and are constantly being improved, but they are focused on the average user and take into account the commercial component. Given this, the search for scientific information is rather difficult. For machine learning in search engines, it is necessary to have a history of user actions that have many variables such as: geolocation, date and time, device type, personalization data, keywords and more. Also a necessary component is to understand the query context and motivation, to understand what the user means. Taking this into account, for the distance learning system, which has become even more relevant in the context of a pandemic, an intelligent search for scientific material has been developed, depending on the interests of the user. The search algorithm forms a group of relevant interests depending on previous search queries and the history of open tabs and forms the search result in scientific journals. Each article contains keywords, so the interest group is formed from the keywords of the articles that interest the user. To check the relevance of user interest groups, the algorithm checks the current search queries for compliance with past interest groups, if the user's interests have changed, forms a new interest group for the user. Forms its knowledge base (article link and short description) depending on what information from the search result was useful to the user, and displays these results to other users on a similar search query. To solve these problems, the existing database of the remote system was expanded.