ALGA: Adaptive lexicon learning using genetic algorithm for sentiment analysis of microblogs

Abstract

Sentiment analysis is about classifying opinions expressed in text. The aim of this study is to improve polarity classification of sentiments in microblogs by building adaptive sentiment lexicons. In the proposed method, corpora-based and lexicon-based approaches are combined and lexicons are generated from text. The sentiment classification is formulated as an optimization problem, in which the goal is to find optimum sentiment lexicons. A novel genetic algorithm is then proposed to solve this optimization problem and find lexicons to classify text. The algorithm generates adaptive sentiment lexicons, and then a meta-level feature is extracted based on it, which is then used alongside Bing Liu's lexicon and n-gram features. The experiments are conducted on six datasets. In terms of accuracy, the results outperform the state-of-the-art methods proposed in the literature in two of the datasets. Also, in four of the datasets, the proposed approach outperforms in terms of Fmeasure. Applying the proposed method on six datasets, the accuracy is higher than 80% in all six datasets and the F-measure is higher than 80% in four of these datasets. Using the sentiment lexicons created by the proposed algorithm, one can get a better understanding of the specific language and culture of Twitter users and sentiment orientation of words in different contexts. It is

Systematic literature review of sentiment analysis on Twitter using soft computing techniques

“…The various benchmark datasets used in the past decade were WePS-3, 27 SemEval, 30,52,54,55,73,75,76,85 tweets prepared by Stanford University, 34,45,46,75 SNAP, 40 Sanders Twitter Sentiment Corpus (denoted as Sanders), 44,55,75,79 2008 Presidential Debate Corpus, 44,75,79 Sentiment140, 51 RepLab 2012, 53 RepLab 2013, 53 STS-manual, 55 Gold Standard personality labeled Twitter dataset, 59 Cleveland Heart Disease data, 69 STS-Gold, 73 FIGURE 6 Distribution of papers in accordance to the digital libraries (expressed in percentages) Many reported researches were carried on the tweets fetched directly from Twitter using its API. The tweets were from a variety of domains, topics and time period (referred as topic specific/topic oriented tweets).…”

“…The tweets were from a variety of domains, topics and time period (referred as topic specific/topic oriented tweets). These prominently included tweets from or about elite personalities like actors; singers; sportsperson; comedians; politicians, authors, idols; entertainers, 28,34,37,40,54,55,67,70,73,75,79 etc, news and commemoratives, 17,30,48,58,59,64,67,79 health and fitness, 31,56,57,69,74,75,78,79 stock market exchanges, 29,34,63,82 companies like AT&T; Amazon;…”

Perspectives of the performance metrics in lexicon and hybrid based approaches: a review

“…A methodology proposed for corpus and lexicon based approaches mainly is used to create the text documents. This approach is used for classifying sentiments and polarity [27]. A genetic algorithm is proposed for optimization problem and is used for finding lexicons in the opinionated text.…”

“…An approach for sentiment classification in micro blogs is proposed. Genetic algorithm, sentiment lexicon, meta-level features, Bing Liu's lexicon and n-gram features are used in the framework [27]. It requires concentrating on creation of the lexicon to reduce the time-consuming and sentiment score in other domains.…”

Hybrid optimization for feature selection in opinion mining

“…Results of the research demonstrated that is improved efficiency of feature selection of classifiers in the classification of opinions. A novel GA was proposed by Keshavarz&Abadeh [13] in solving optimization issues and to find lexicon to classify text. Adaptive sentiment lexicons were generated through this algorithm and on its basis, which was used along with Bing Liu's lexicon and ngram features.…”