Optimizing the Hyperparameter of Feature Extraction and Machine Learning Classification Algorithms
Journal Title: International Journal of Advanced Computer Science & Applications - Year 2019, Vol 10, Issue 3
Abstract
The process of assigning a quantitative value to a piece of text expressing a mood or effect is called Sentiment analysis. Comparison of several machine learning, feature extraction approaches, and parameter optimization was done to achieve the best accuracy. This paper proposes an approach to extracting comparison value of sentiment review using three features extraction: Word2vec, Doc2vec, Terms Frequency-Inverse Document Frequency (TF-IDF) with machine learning classification algorithms, such as Support Vector Machine (SVM), Naive Bayes and Decision Tree. Grid search algorithm is used to optimize the feature extraction and classifier parameter. The performance of these classification algorithms is evaluated based on accuracy. The approach that is used in this research succeeded to increase the classification accuracy for all feature extractions and classifiers using grid search hyperparameter optimization on varied pre-processed data.
Authors and Affiliations
Sani Muhammad Isa, Rizaldi Suwandi, Yosefina Pricilia Andrean
Keywords
Related Articles
- EP ID EP498383
- DOI 10.14569/IJACSA.2019.0100309
- Views 100
- Downloads 0
Electrooculogram Signals Analysis for Process Control Operator Based on Fuzzy c-Means
Biomedical signals of human can reflect the body's task load, fatigue and other psychological information. Compared with other biomedical signals, electrooculogram (EOG) has higher amplitude, less interference, and is ea...
Comparing the Usability of M-Business and M-Government Software in Saudi Arabia
This study presents a usability assessment of mobile presence in the Kingdom of Saudi Arabia (KSA), with a particular focus on the variance between M-business and M-government presence. In fact, a general hypothesis was...
An Efficient Design of RPL Objective Function for Routing in Internet of Things using Fuzzy Logic
The nature of the Low power and lossy networks (LLNs) requires having efficient protocols capable of handling the resource constraints. LLNs consist of networks that connect different type of devices which has constraint...
A Tri-Level Industry-Focused Learning Approach for Software Engineering Management
Most engineering classes in higher education rely heavily on the traditional lecture format, despite the fact that a number of investigations have shown that lectures, even when given by good lecturers, have limited succ...
A Survey of Datasets for Biomedical Question Answering Systems
The massively ever increasing amount of textual and linked biomedical data available online poses many challenges for information seekers. So, the focus of information retrieval community has shifted to precise informati...