Sentiment Analysis using Machine Learning (The Sorting Hat)

Authors

  • Prasad Wagh
  • Pratik Jaiswal
  • Ankit Rahangdale

DOI:

https://doi.org/10.46243/jst.2021.v6.i04.pp377-382

Keywords:

Sorting Hat, Hate Speech, Sarcasm, Opinion Mining, Sentiment Analysis

Abstract

Sarcasm and Hate Speech is impacting societal harmony and peace. Considering the magnitude of this harmonious impact, there is a need to find a solution to curb the online spread of Hate Speech and Sarcasm. Detection of hate speech and sarcasm is being tackled with various approaches like manual checks, deep learning techniques in recent times, and statistical-based classification algorithms. These methods are unreliable due to the non-binary(true or false) nature of the tweets. Categorizing tweets requires deeper investigation such as classification on entirely positive or entirely negative rather than binary classification. In this paper - a snippet - The Sorting Hat, to detect sarcasm, hate-speech, and sentiments in the tweets using SVM (Support Vector Machine) and LSTM (Long short term memory) is proposed. The Sorting Hat classifies a given tweet into one of the six degrees of classification - “Positive”, “Negative”, “Neutral”, “Sarcasm”, “Non-sarcasm”, “Hate-speech”. The basic meaning of sarcasm which comes into our mind is a positive statement or sentiment attached to a negative situation or vice versa.The current system works on the outside whiсh has been assigned а раrtiсulаr tорiс. Current systems also do not determine the imрасt rating, the results are limited to whether they can be included in the раrtiсulаr processing field and do not allow retrieval of data based on user-generated query meaning that it has been selected.. Whereas the Sorting Hat will collect the tweets from the users manually. Collected tweets will be considered for further processing. We will then аррly the suрervised аlgоrithm оn the stоred dаtа. The supervised algorithm used in the оur system is Suрроrt Veсtоr Mасhine (SVM). The results of the algorithms i.e. emotions will be represented in a graphical way (bar charts). The proposed system works better compared to the existing one. This is because we will be able to obtain calculated figures from reрresentаtiоn оf result саn hаvе аny imрасt in the field of а раrtiсulаr. The overall product experience using The Sorting Hat largely intervenes the impulsive behavior of posting tweets, and thereby provides the solution to curb rampant spread of Hate Speech and better understanding of sarcastic tweets.

 

 

Downloads

Published

2021-08-16

How to Cite

Prasad Wagh, Pratik Jaiswal, & Ankit Rahangdale. (2021). Sentiment Analysis using Machine Learning (The Sorting Hat). Journal of Science & Technology (JST), 6(Special Issue 1), 377–382. https://doi.org/10.46243/jst.2021.v6.i04.pp377-382