Abstract
Opinion mining and sentiment analysis are aimed at extracting valuable necessary information from the raw text and determining the polarity of the text. Comparative opinion mining analyzes opinions that compare two or more entities based on a common feature. The purpose of comparative opinions mining is to extract comparative opinions from the text and identify the objects being compared. First step of identification comparative opinions is extraction comparative constructions. There are presently research papers on sentiment analysis and opinion mining in the Kazakh language. But we were unable to find any work on the comparative opinion mining. The aim of the study is to automate the extraction of comparative constructions in the Kazakh language. In this paper, a pattern based approach was implemented to extract comparative constructions in the Kazakh language.