Search Query Classification Using Machine Learning for Information Retrieval Systems in Intelligent Manufacturing
Year: 2019
Type: conference-paper
Abstract: The problem of the information retrieval systems operation efficiency increasing in manufacturing is analysed and the general statement of research task is carried out. The training sample preparation for search queries classification was performed, during which the task of search queries classification was formulated, pre-processing of the search query texts was carried out and the dictionary was generated. For the dictionary formation, search queries tokenization and stamping were initially carried out, and then a dictionary was created in the words weight vector form based on the search query text generalized evaluation. On the basis of prepared dictionary (as the matrix of 121×1119 elements) the search queries classifier model was developed and experimentally investigated. The neural network learning was performed. In the result the network with best performance was determined, for which neurons number for the first layer is 30, and for second layer is 5. The developed neural network can be integrated into the information retrieval system in the form of a classifier in 5 categories. It will make possible to classify search queries and thereby increase the information retrieval system operation speed in manufacturing.
Institution: Kharkiv National University of Radio Electronics
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W2955152148
