Feature Selection Methods in Sentiment Analysis

Khairi, Nurilhami Izzatie; Mohamed, Azlinah; Yusof, Nor Nadiah

doi:10.1145/3386723.3387840

Cited by 3 publications

(4 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…trigrams) to produce higher classification accuracy with the context of terms (words) in mind [16]. A study by [34] found that hybrid methods achieved the best CPU performance and accuracy levels even on feature selection. In the subsequent sub-sections, we summarize the machine learning techniques and a popular deep learning technique we used in developing our hybrid machine learning techniques for experimentation.…”

Section: Feature Engineeringmentioning

confidence: 99%

Hybrid Machine Learning Techniques for Comparative Opinion Mining

Ondara,

Waithaka,

Kandiri

et al. 2023

IJAIDM

View full text Add to dashboard Cite

Comparative opinion mining has lately gained traction among individuals and businesses due to its growing range of applications in brand reputation monitoring and consumer decision making among others. Past research in sub-field of opinion mining have mostly explored single-entity opinion mining models and the mining of comparative sentences suing single classifiers. Most of these studies relied on a limited number of comparative opinion labels and datasets while applying the techniques in limited domains. Consequently, the reported performances of the techniques might not be optimal in some cases like working with big data. In this study, however, we developed four hybrid machine learning techniques, with which we performed multi-class based comparative opinion mining using three datasets from different domains. From our results, the best-performing hybrid machine learning technique for comparative opinion mining using a multi-layer perceptron as the base estimator was the Multilayer Perceptron + Random Forest (MLP + RF). This technique had an average accuracy of 93.0% and an F1-score of 93.0%. These results show that our hybrid machine learning techniques could reliably be used for comparative opinion mining to support business needs like brand reputation monitoring.

show abstract

Section: Feature Engineeringmentioning

confidence: 99%

Hybrid Machine Learning Techniques for Comparative Opinion Mining

Ondara,

Waithaka,

Kandiri

et al. 2023

IJAIDM

View full text Add to dashboard Cite

show abstract

“…Lexicon-based approaches can be categorized as ones of pure NLP techniques and these approaches have gained popularity in the sentiment analysis field, especially during 2016-2017. Meanwhile, the recent advances in the sentiment analysis field have seen the emergence of new techniques that are shifted from pure NLP to conventional machine learning classifiers, as well as deep learning-based models that have been shown to outperformed other methods by achieving state-of-the-art performances [23].…”

Section: Natural Language Processing Overviewmentioning

confidence: 99%

“…Document level aims to identify the sentiment of the whole document, meanwhile sentence level is more fine-grained and aims to analyze the sentiment of an individual sentence. Another level of sentiment analysis is the aspect level, which aims to identify the sentiment of an attribute that are usually being discussed together during a product review process [23].…”

Section: Sentiment Analysis On Twittermentioning

confidence: 99%

Climate Change Sentiment Analysis Using Lexicon, Machine Learning and Hybrid Approaches

Sham

Mohamed

2022

Sustainability

Self Cite

View full text Add to dashboard Cite

The emissions of greenhouse gases, such as carbon dioxide, into the biosphere have the consequence of warming up the planet, hence the existence of climate change. Sentiment analysis has been a popular subject and there has been a plethora of research conducted in this area in recent decades, typically on social media platforms such as Twitter, due to the proliferation of data generated today during discussions on climate change. However, there is not much research on the performances of different sentiment analysis approaches using lexicon, machine learning and hybrid methods, particularly within this domain-specific sentiment. This study aims to find the most effective sentiment analysis approach for climate change tweets and related domains by performing a comparative evaluation of various sentiment analysis approaches. In this context, seven lexicon-based approaches were used, namely SentiWordNet, TextBlob, VADER, SentiStrength, Hu and Liu, MPQA, and WKWSCI. Meanwhile, three machine learning classifiers were used, namely Support Vector Machine, Naïve Bayes, and Logistic Regression, by using two feature extraction techniques, which were Bag-of-Words and TF–IDF. Next, the hybridization between lexicon-based and machine learning-based approaches was performed. The results indicate that the hybrid method outperformed the other two approaches, with hybrid TextBlob and Logistic Regression achieving an F1-score of 75.3%; thus, this has been chosen as the most effective approach. This study also found that lemmatization improved the accuracy of machine learning and hybrid approaches by 1.6%. Meanwhile, the TF–IDF feature extraction technique was slightly better than BoW by increasing the accuracy of the Logistic Regression classifier by 0.6%. However, TF–IDF and BoW had an identical effect on SVM and NB. Future works will include investigating the suitability of deep learning approaches toward this domain-specific sentiment on social media platforms.

show abstract

“…For estimating variance of SNPs, comprising values of a gene related to prone or salubrious records of the chosen training set, the method adapts ANOVA standard bidirectional pooled variance estimation. Based on results envisaged in [30,31], the method is chosen for analysis. The bidirectional pooled variance estimation is adapted for selecting optimal features pertaining to every record (both prone and salubrious) for a training set chosen.…”

Section: Bidirectional Pooled Variance Estimationmentioning

confidence: 99%

Genome feature optimization and coronary artery disease prediction using cuckoo search

Neelima,

Babu

2020

Comput Sci Inf Technol

View full text Add to dashboard Cite

Cardiovascular diseases are among the major health ailment issue leading to millions of deaths every year. In recent past, analyzing gene expression data, particularly using machine learning strategies to predict and classify the given unlabeled gene expression record is a generous research issue. Concerning this, a substantial requirement is feature optimization, which is since the overall genes observed in human body are closely 25000 and among them 636 are cardiovascular related genes. Hence, it complexes the process of training the machine learning models using these entire cardiovascular gene features. This manuscript uses bidirectional pooled variance strategy of ANOVA standard to select optimal features. Along the side to surpass the constraint observed in traditional classifiers, which is unstable accuracy at k-fold cross validation, this manuscript proposed a classification strategy that build upon the swarm intelligence technique called cuckoo search. The experimental study indicating that the number of optimal features those selected by proposed model is substantially low that compared to the other contemporary model that selects features using forward feature selection and classifies using support vector machine classifier (FFS&SVM). The experimental study evinced that the proposed model, which selects feature by bidirectional pooled variance estimation and classifies using proposed classification strategy that build on cuckoo search (BPVE&CS) outperformed the selected contemporary model (FFS&SVM).

show abstract

Feature Selection Methods in Sentiment Analysis

Cited by 3 publications

References 29 publications

Hybrid Machine Learning Techniques for Comparative Opinion Mining

Hybrid Machine Learning Techniques for Comparative Opinion Mining

Climate Change Sentiment Analysis Using Lexicon, Machine Learning and Hybrid Approaches

Genome feature optimization and coronary artery disease prediction using cuckoo search

Contact Info

Product

Resources

About