RETRACTED ARTICLE: Topic flexible aspect based sentiment analysis using minimum spanning tree with Cuckoo search
| Autor: | I. Mohan, M. Moorthi |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
General Computer Science
Computer science business.industry Sentiment analysis Feature selection 02 engineering and technology Minimum spanning tree Machine learning computer.software_genre Random forest Identification (information) 020204 information systems 0202 electrical engineering electronic engineering information engineering Feature (machine learning) 020201 artificial intelligence & image processing AdaBoost Artificial intelligence Cuckoo search business computer |
| Zdroj: | Journal of Ambient Intelligence and Humanized Computing. 12:7399-7406 |
| ISSN: | 1868-5145 1868-5137 |
| Popis: | A study by which the opinions, emotions, evaluations, appraisals, and sentiments of people towards different entities is expressed in the form of a text is called Sentiment Analysis (SA). The primary task of a sentiment analysis which is based on the aspects is the extraction of various aspects of the entities and the determination of the sentiments that correspond to the terms of aspects which are commented in the review document. Recently, there is a huge rise in interest to make an identification of various sentiments and aspects at the same time. Feature selection in terms of aspects of entity plays a crucial role in deciding the efficiency of the sentiment analysis; hence the Minimum Spanning Tree (MST) is used for feature selection.The MST has certain major advantages such as being computable quickly. The selection of optimal features to aid in better accuracy of classification is done through MST optimized with Cuckoo search algorithm. The features in sentiment analysis are classified using Random Forest (RF) and Ada Boost classifiers.The Random Forest (RF) is probably the most accurate among all algorithms of learning available today. The Ada Boost algorithm has a performance that is extremely good owing to its ability to be able to generate the expanding diversity. This was done in order to bring about an improvement in the final ensemble, as it contained several weak classifiers. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink