International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 08 | Aug 2024
p-ISSN: 2395-0072
www.irjet.net
Machine Learning vs Deep Learning Approach for Sentiment Analysis on Twitter Data Swapnil sonawane, Alisha Gaikwad, Sneha Thakur ------------------------------------------------------------------------***-----------------------------------------------------------------------
Abstract- Sentiment Analysis is a process of categorizing
like elections, and wars to keep the platform safe and make sure that the platform stays neutral for all users and does not get biased in a specific direction. These companies also used sentiment analysis to monitor the tweets and posts to make sure that they are appropriate and follow all the community guidelines, and if someone goes against the guidelines, they remove their content from the platform. We can use Machine Learning and Deep Learning algorithms to classify the sentiment. However, which technique is suitable for problem statements, depends on the data size and ability to adapt to new contexts.
whether the text is positive, negative, or neutral. Not only this, but it also includes emotions like happiness, sadness, anger, fear, and surprise. Sentiment analysis can be used in various fields, some well-known fields including online shopping. Sentiment analysis can solve real-time issues and is a crucial task in Natural Language Processing (NLP). We can use traditional Machine Learning algorithms, such as Support Vector Machine (SVM), Tree-based techniques, or Naïve Bayes have been widely used for sentiment analysis. The advent of deep learning (DL) techniques, like CNN, RNN, or state-of-the-art methods, changed this field to capture the more complex patterns in data. This paper presents the comparative study of sentiment analysis using ML and DL techniques. We used ML and DL algorithms against Twitter Sentiment Analysis data and compared the algorithms based on accuracy, computational efficiency, and ability to adopt complex patterns in large datasets. This paper provides insights into the trade-off between ML and DL approaches for Sentiment Analysis, further guiding researchers and practitioners in choosing the appropriate approach for their specific tasks.
There are a few observations on which the system architecture is proposed.
2. LITERATURE REVIEW 2.1 Lexical or Rule-Based Approach In “Twitter Sentiment Analysis Using Lexical or RuleBased Approach: A Case Study” [1], Sheresh Zahoor and Rajesh Rohila use Lexical or Rule Based (unsupervised technique) for Twitter sentiment analysis. Using the Twitter API, they create 4 different datasets. 1. Haryana Assembly Polls 2. ML Khattar 3. The sky is pink (movie) 4. United Nations General Assembly (UNGA). The steps they follow to collect the data and analyze the sentiments are:
Key Words: Sentiment Analysis, Machine Learning, Deep Learning, RNN, LSTM, GRU
1. INTRODUCTION
1. 2. 3. 4.
As the whole world connects to the internet, data is everywhere, and the famous quote “Data is the new Oil” is relevant to the current world. This data includes customer feedback, reviews on the products they buy, and people’s opinions on various topics on different social media platforms. E-commerce websites like Amazon, Flipkart, and Walmart, must analyze the customer's feedback and review the product to increase the product sales. Also, the manufacturer can improve and address the customer's concerns, to enhance the customer's experience and satisfaction by analyzing the sentiment of customers' reviews and feedback. Sentiment analysis can help businesses to monitor their reputation by tracking the comments, and social media reviews. By understanding the user's review, product developer can improve their products. Companies like Twitter, Facebook, and Instagram can analyze people’s opinions on current trending topics
© 2024, IRJET
|
Impact Factor value: 8.315
Data Collection Data pre-processing Part of Speech tagging (POS) Sentiment analysis using an in-built dictionary
A. Data Collection: To collect data from Twitter, they use the Twitter API, collect the tweet, and save it in CSV format. The CSV file contains the date, text, retweet, hashtag, and followers. B. Data Pre-Processing: To prepare data for sentiment analysis, they perform various operations on data, including tokenization or Bag-of-words, N-gram Extraction, Stemming and Lemmatization, and StopWords removal.
|
ISO 9001:2008 Certified Journal
|
Page 636