Deep learning model for Mongolian Citizens Feedback Analysis using Word Vector Embeddings

TL;DR

This paper develops deep learning models using custom word embeddings to analyze Mongolian citizen feedback, achieving over 80% accuracy, addressing language-specific NLP challenges.

Contribution

It introduces a feedback classification approach for Mongolian using two different word embeddings and compares their effectiveness.

Findings

Word embeddings trained on Mongolian data improve model accuracy.

Achieved up to 82.7% accuracy in feedback classification.

Demonstrated feasibility of deep learning for low-resource language NLP.

Abstract

A large amount of feedback was collected over the years. Many feedback analysis models have been developed focusing on the English language. Recognizing the concept of feedback is challenging and crucial in languages which do not have applicable corpus and tools employed in Natural Language Processing (i.e., vocabulary corpus, sentence structure rules, etc). However, in this paper, we study a feedback classification in Mongolian language using two different word embeddings for deep learning. We compare the results of proposed approaches. We use feedback data in Cyrillic collected from 2012-2018. The result indicates that word embeddings using their own dataset improve the deep learning based proposed model with the best accuracy of 80.1% and 82.7% for two classification tasks.