Penguin Search Optimization Based Feature Selection for Automated Opinion Mining

Abstract

Twitter sentiment analysis is a vital concept in determining the public opinions about products, services, events or personality. Analyzing the medical tweets on a specific topic can provide immense benefits in medical industry. However, the medical tweets require efficient feature selection approach to produce significantly accurate results. Penguin search optimization algorithm (PeSOA) has the ability to resolve NP-hard problems. This paper aims at developing an automated opinion mining framework by modeling the feature selection problem as NP-hard optimization problem and using PeSOA based feature selection approach to solve it. Initially, the medical tweets based on cancer and drugs keywords are extracted and pre-processed to filter the relevant informative tweets. Then the features are extracted based on the Natural Language Processing (NLP) concepts and the optimal features are selected using PeSOA whose results are fed as input to three baseline classifiers to achieve optimal and accurate sentiment classification. The experimental results obtained through MATLAB simulations on cancer and drug tweets using k-Nearest Neighbor (KNN), Naïve Bayes (NB) and Support Vector Machine (SVM) indicate that the proposed PeSOA feature selection based tweet opinion mining has improved the classification performance significantly. It shows that the PeSOA feature selection with the SVM classifier provides superior sentiment classification than the other classifiers

DEA-RNN: A Hybrid Deep Learning Approach for Cyberbullying Detection in Twitter Social Media Platform

“…The features from the Twitter dataset are extracted using NLP tools such as Word2Vec and TF-IDF, with the nouns, pronouns, and adjectives are considered as primary feature contents, whereas the adverbs and verbs provide additional information. Furthermore, the extraction of Part-of-Speech (POS) tags, function words, and content word features can improve the classification performance [60]. For identifying the cyber-bullying events, prominent feature are selected utilizing the Information Gain (GI) method, then these features subsets are fed into DEA-RNN classifier.…”

Feature Selection Problem and Metaheuristics: A Systematic Literature Review about Its Formulation, Evaluation and Applications

“…It combines precision and recall into a single value and is particularly useful for balancing these metrics. This metric has been used in [22,33,35,39,45,48,51,55,59,60,83,92,98,101,117,132,141,149,155,162,164,167,171,[176][177][178]182,184,185] and mathematically is defined as follows:…”

“…Measures the model's ability to correctly identify all positive instances. This metric has been used in [22,33,35,55,56,60,70,85,96,98,101,102,106,117,129,132,133,141,149,162,164,171,[176][177][178]185] and mathematically is defined as follows:…”

A Review of Feature Selection Algorithms in Sentiment Analysis for Drug Reviews

“…NLP concepts, such as part of speech tagging, n-gram, content words, and function words have also been used to extract features from tweet data [16]. The Penguin Search Optimization (PeSOA) algorithm [16] was also used as a feature selection technique to select optimal features based on the keywords of drugs and cancer in tweet data.…”

“…NLP concepts, such as part of speech tagging, n-gram, content words, and function words have also been used to extract features from tweet data [16]. The Penguin Search Optimization (PeSOA) algorithm [16] was also used as a feature selection technique to select optimal features based on the keywords of drugs and cancer in tweet data. The K-Nearest Neighbour (KNN), Naïve Bayes (NB), and Support Vector Machine (SVM) methods were used through MATLAB simulation software to classify the tweet data.…”