Розробка експертної системи аналізу результатів вступних компаній до ВНЗ

Abstract

Аналіз мережі Інтернет на наявність великих обсягів даних про студентів спеціальності "комп'ютерна інженерія". Синтаксичний аналіз (парсинг) сайтів https://abit-poisk.org.ua і https://vk.com, з використанням мови програмування Python, за допомогою фреймворка Scrapy і vk API. Статистичний аналіз отриманих даних за методом найменших квадратів. Побудова моделей машинного навчання за методами KNN і Random Forest.Analysis of the Internet for the availability of large volumes of data on students specialty "computer engineering". Parsing (parsing) sites https://abit-poisk.org.ua and https://vk.com, using the Python programming language, using the Scrapy framework and the vk API. Statistical and intellectual analysis of collected data to improve the work of the UDCTU receiving commission, using the least squares method. Construction of models of machine learning using KNN and Random Forest methods. Obtained results: the largest mean values of the passing ball before joining is shown by the Department of DNU them. O. Gonchar, this can be explained by more budget places; regarding the maximum and minimum average scores, then the approximate picture for all departments of the city is approximately the same; the left bank of the city of Dnipro is an area that provided the minimum number of students who have passed the competitive selection on a specialty; unlike the previous point, the city center produces a large number of potential specialists. During the study, the means of collecting information on the Internet, methods for analyzing big data and ways of constructing data models with the help of machine learning and neural networks are analyzed and studied. On the basis of the collected information, an expert system was created that provides information gathering for entrants on the basis of which statistical and intellectual analysis was conducted. In the future, you can develop an expert system in several directions: a more detailed analysis of each prospective student, with a psychological portrait on behavior in the Internet;tracking the success of students; drawing up a list of persons for targeting advertising in social networks.