Multi-level particle swarm optimisation and its parallel version for parameter optimisation of ensemble models: a case o
- PDF / 909,314 Bytes
- 16 Pages / 595.276 x 790.866 pts Page_size
- 80 Downloads / 249 Views
(0123456789().,-volV)(0123456789(). ,- volV)
Multi-level particle swarm optimisation and its parallel version for parameter optimisation of ensemble models: a case of sentiment polarity prediction Gregorius Satia Budhi1,2 • Raymond Chiong1 • Sandeep Dhakal1 Received: 9 September 2019 / Revised: 8 January 2020 / Accepted: 9 March 2020 Ó Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Ensemble learning is increasingly used in sentiment analysis. Determining the parameter settings of ensemble models, however, is not easy. Besides its own parameters, an ensemble model has base-predictors that have their individual parameters. Some ensemble models use a specific base-predictor and could be optimised using standard metaheuristics such as the Particle Swarm Optimisation (PSO) approach. Optimising ensemble models with multiple base-predictor candidates is more complicated and challenging, as there are multiple options to choose from. We therefore propose MultiLevel PSO (ML-PSO) and Parallel ML-PSO (PML-PSO) to optimise the parameters of ensemble models, especially those with multiple base-predictors, for sentiment analysis. The idea is to utilise multiple PSOs as particles of the main PSO. The main PSO optimises ensemble-model parameters and determines the best base-predictor, whereas PSOs within it optimise the corresponding base-predictor’s parameters. Experimental results using Bagging Predictors as the underlying ensemble model show that ML-PSO can improve prediction accuracy, while PML-PSO is able to speed up the processing time and further improve the accuracy. Keywords Particle swarm optimisation Parallelism Machine learning Sentiment analysis
1 Introduction Sentiment polarity detection, or more generally known as sentiment analysis, is the process of automatically and systematically detecting the sentiment or opinion of a given text. In addition to feature selection, the outcome of sentiment analysis primarily depends on the detection algorithm applied [1–4]. The majority of methods used for sentiment analysis belong to the machine learning domain. These methods are usually applied to predict the sentiment polarity of social media texts, online product reviews or other kinds of texts [2–8]. Due to the extensive amount of online texts such as product reviews, tweets and other
& Raymond Chiong [email protected] 1
School of Electrical Engineering and Computing, The University of Newcastle, Callaghan, NSW 2308, Australia
2
Informatics Department, Petra Christian University, Surabaya 60236, Indonesia
social media texts, a system capable of automated sentiment analysis is vital in the online environment [9, 10]. Analysis using machine learning generally begins with training the machines to make them capable of discriminating the texts. The accuracy of the prediction model is determined by the quality of this training process [4, 11], and also how features of the text are extracted [12, 13]. However, acquiring the correct parameter settings for machine learning model
Data Loading...
