An end-to-end ranking system based on customers reviews: Integrating semantic mining and MCDM techniques
Welcome. This repository contains the python based implementation of the 'An end-to-end ranking system based on customers reviews: Integrating semantic mining and MCDM techniques'. In this repository, source codes of our paper are presented in terms of web scraping, offline and online phases. In corresponded paper, we propose an end-to-end ranking method for integrating mechanisms such as text processing, sentiment analysis and the multi-criteria decision-making technique. The proposed ranking method relies on the integration of three methods, namely, the aspect-based sentiment analysis (ABSA) method, the Dawid-Skene algorithm and the Best Worst Method (BWM). In other words, the proposed work encompasses four major steps: i) crawling customer reviews, ii) preprocessing, iii) aspect term extraction, aspect category detection and polarity detection, and iv) designing a decision-making model.
In this repository, there are three folders namely #Scrap, #online and #offline phases. In each folder implementation of correspond phase is presented. note that before run codes, you must download Glove 6b.200 pretrained word2wec model from https://nlp.stanford.edu/projects/glove/ and paste it in the data folder which exist in online and offline folders. Also, crawled data which are collected from https://www.tripadvisor.com are existe in crawled_data folder. This data has been collected using codes exist in #Scrape folder and due to possible changes in the site structure, they should be changed according to the latest site changes before run.
If you find this code useful please cite us in your work:
Milad Eshkevari, Mustafa Jahangoshai Rezaee, Morteza Saberi, Omar K. Hussain, An end-to-end ranking system based on customers reviews: Integrating semantic mining and MCDM techniques, Expert Systems with Applications, Volume 209, 2022, 118294, ISSN 0957-4174, https://doi.org/10.1016/j.eswa.2022.118294. (https://www.sciencedirect.com/science/article/pii/S0957417422014294)