Garuda - Garba Rujukan Digital

Sentiment Analysis on Song Lyrics for Song Popularity Prediction Using BERT Agatha, Hana; Putri, Farica Perdana; Suryadibrata, Alethea
ULTIMATICS Vol 15 No 2 (2023): Ultimatics : Jurnal Teknik Informatika
Publisher : Faculty of Engineering and Informatics, Universitas Multimedia Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31937/ti.v15i2.3420

The increasingly competitiveness in music industry is giving some musicians disadvantages. Musicians need to pay more attention to the factors that influence the popularity of a song, so their song can be popular and they can gain a lot of profit. One of the various factors that can affect a song popularity is the lyrics. The influence of the lyrics can be explored through sentiment analysis. Sentiment analysis is a computing study that identify sentiments or emotions in a text. By conducting sentiment analysis on the lyrics, song popularity can be predicted. Based on the prediction result, songwriters can evaluate their lyrics, so their song can be popular. Bidirectional Encoder Representations from Transformers (BERT) is an excellent algorithm in terms of sentiment analysis. In this study, a BERT model was developed to predict the song popularity, based on the sentiment analysis of the song lyrics. The popularity class of a song will be predicted, based on the results of lyrics sentiment analysis. The developed model is a model that has been trained with English songs. Based on the experiment, the model that used oversampling method achieved accuracy by 87%, precision by 88%, recall by 87%, and f1-score by 87%.

Optimized support vector machine for sentiment analysis of game reviews Supriyatna, Bryan Leonardo; Putri, Farica Perdana
International Journal of Informatics and Communication Technology (IJ-ICT) Vol 13, No 3: December 2024
Publisher : Institute of Advanced Engineering and Science

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/ijict.v13i3.pp344-353

The rapid development of games has made game categories diverse, so there are many opinions about games that have been released. Sentiment analysis on game reviews is needed to attract potential players. Sentiment analysis is carried out using the support vector machine (SVM) and particle swarm optimization (PSO) algorithms. SVM training was conducted with a linear kernel, the ‘C’ value parameter was 10 resulting in an accuracy value of 97.28%. The SVM algorithm optimized using the PSO method produces an accuracy of 97.61% using the parameters c1 is 0.2, c2 is 0.5 and w is 0.6. Based on these results, sentiment analysis using PSO-based SVM optimization has been successfully carried out with an increase in accuracy of 0.33%. This game review has a sentiment value from neutral to positive so this game can be recommended to other players.

Improved SVM for Website Phishing Detection Through Recursive Feature Elimination Putri, Farica Perdana; Marcello, Feliciano Surya
Ultimatics : Jurnal Teknik Informatika Vol 16 No 2 (2024): Ultimatics : Jurnal Teknik Informatika
Publisher : Faculty of Engineering and Informatics, Universitas Multimedia Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31937/ti.v16i2.3744

Technology is developing faster every day, particularly in the information technology field. A website is one of the many information access points people use to do business activities, get information, and other purposes. Sophisticated websites are being developed and used, encouraging many naive individuals to commit crimes for financial gain. Phishing websites are a common method of using information technology to conduct fraud. One way to conduct phishing is by using the features on the website. One technique for identifying phishing websites is to use the Support Vector Machine (SVM) algorithm, which classifies websites based on features. However, the SVM algorithm is not able to detect many features so that the resulting accuracy and optimization level is also not good. Based on datasets, the SVM algorithm only gets around 60% to 70% accuracy. The use of Recursive Feature Elimination (RFE) feature selection is one way that can be done to cover the shortcomings of SVM. By eliminating features that irrelevant and redundance, RFE makes the SVM algorithm get a higher accuracy rate on the available dataset with an accuracy of 96.09%.

Cross-Platform Mobile Based Crowdsourcing Application for Sentiment Labeling Using Gamification Method Elaine, Elaine; Putri, Farica Perdana; Suryadibrata, Alethea
IJNMT (International Journal of New Media Technology) Vol 11 No 2 (2024): Vol 11 No 2 (2024): IJNMT (International Journal of New Media Technology)
Publisher : Universitas Multimedia Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31937/ijnmt.v11i2.3935

Sentiment analysis is the application of natural language processing which aims to identify the sentiment of texts. To carry out sentiment analysis, data which has been labeled sentiment is needed to be included in the training model. Crowdsourcing is considered as the most optimal method to label data because it has a high level of accuracy at a relatively low cost. However, the use of crowdsourcing platforms has its own challenge, which is to increase user interest and motivation. A solution which can be applied is to design and build a crowdsourcing platform or application using the gamification method. The definition of gamification is an effort to increase one's intrinsic motivation for an activity by applying game elements to it. Therefore, a cross-platform mobile based crowdsourcing application for sentiment labeling using gamification method was carried out. The gamification design process was done based on the 6D framework and the application was developed using the Ionic-React framework. Application was examined through black box testing and the result showed that the application was functioning properly and according to the design requirements. There was also an evaluation carried out by distributing Intrinsic Motivation Inventory questionnaires to users who had used the application for 2 weeks. From a total of 40 respondents, the result showed that the level of user motivation and interest in using the application was high with a percentage of 83.10%.

Cross-Platform Mobile Based Crowdsourcing Application for Sentiment Labeling Using Gamification Method Elaine, Elaine; Putri, Farica Perdana; Suryadibrata, Alethea
IJNMT (International Journal of New Media Technology) Vol 11 No 2 (2024): Vol 11 No 2 (2024): IJNMT (International Journal of New Media Technology)
Publisher : Universitas Multimedia Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31937/ijnmt.v11i2.3935

Sentiment analysis is the application of natural language processing which aims to identify the sentiment of texts. To carry out sentiment analysis, data which has been labeled sentiment is needed to be included in the training model. Crowdsourcing is considered as the most optimal method to label data because it has a high level of accuracy at a relatively low cost. However, the use of crowdsourcing platforms has its own challenge, which is to increase user interest and motivation. A solution which can be applied is to design and build a crowdsourcing platform or application using the gamification method. The definition of gamification is an effort to increase one's intrinsic motivation for an activity by applying game elements to it. Therefore, a cross-platform mobile based crowdsourcing application for sentiment labeling using gamification method was carried out. The gamification design process was done based on the 6D framework and the application was developed using the Ionic-React framework. Application was examined through black box testing and the result showed that the application was functioning properly and according to the design requirements. There was also an evaluation carried out by distributing Intrinsic Motivation Inventory questionnaires to users who had used the application for 2 weeks. From a total of 40 respondents, the result showed that the level of user motivation and interest in using the application was high with a percentage of 83.10%.

Sentiment Analysis on Song Lyrics for Song Popularity Prediction Using BERT Agatha, Hana; Putri, Farica Perdana; Suryadibrata, Alethea
ULTIMATICS Vol 15 No 2 (2023): Ultimatics : Jurnal Teknik Informatika
Publisher : Faculty of Engineering and Informatics, Universitas Multimedia Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31937/ti.v15i2.3420

The increasingly competitiveness in music industry is giving some musicians disadvantages. Musicians need to pay more attention to the factors that influence the popularity of a song, so their song can be popular and they can gain a lot of profit. One of the various factors that can affect a song popularity is the lyrics. The influence of the lyrics can be explored through sentiment analysis. Sentiment analysis is a computing study that identify sentiments or emotions in a text. By conducting sentiment analysis on the lyrics, song popularity can be predicted. Based on the prediction result, songwriters can evaluate their lyrics, so their song can be popular. Bidirectional Encoder Representations from Transformers (BERT) is an excellent algorithm in terms of sentiment analysis. In this study, a BERT model was developed to predict the song popularity, based on the sentiment analysis of the song lyrics. The popularity class of a song will be predicted, based on the results of lyrics sentiment analysis. The developed model is a model that has been trained with English songs. Based on the experiment, the model that used oversampling method achieved accuracy by 87%, precision by 88%, recall by 87%, and f1-score by 87%.

Improved SVM for Website Phishing Detection Through Recursive Feature Elimination Putri, Farica Perdana; Marcello, Feliciano Surya
ULTIMATICS Vol 16 No 2 (2024): Ultimatics : Jurnal Teknik Informatika
Publisher : Faculty of Engineering and Informatics, Universitas Multimedia Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31937/ti.v16i2.3744

Technology is developing faster every day, particularly in the information technology field. A website is one of the many information access points people use to do business activities, get information, and other purposes. Sophisticated websites are being developed and used, encouraging many naive individuals to commit crimes for financial gain. Phishing websites are a common method of using information technology to conduct fraud. One way to conduct phishing is by using the features on the website. One technique for identifying phishing websites is to use the Support Vector Machine (SVM) algorithm, which classifies websites based on features. However, the SVM algorithm is not able to detect many features so that the resulting accuracy and optimization level is also not good. Based on datasets, the SVM algorithm only gets around 60% to 70% accuracy. The use of Recursive Feature Elimination (RFE) feature selection is one way that can be done to cover the shortcomings of SVM. By eliminating features that irrelevant and redundance, RFE makes the SVM algorithm get a higher accuracy rate on the available dataset with an accuracy of 96.09%.

Title

Found 7 Documents
Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title Search

Found 7 Documents Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title

Found 7 Documents
Search