Indonesian text feature extraction using gibbs sampling and mean variational inference latent dirichlet allocation

Prihatini, Putu Manik; Putra, I Ketut Gede Darma; Giriantari, Ida Ayu Dwi; Sudarma, Made

doi:10.1109/qir.2017.8168448

Cited by 10 publications

(7 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Case folding or lowercasing is the process of converting all characters from the clean text dataset into lowercase [10]. This process is done to avoid errors towards identifying a specific term in the dataset [25].…”

Section: Preprocessingmentioning

confidence: 99%

Hate Speech Classification in Indonesian Language Tweets by Using Convolutional Neural Network

Taradhita

Putra

2021

J. ICT Res. Appl.

View full text Add to dashboard Cite

The rapid development of social media, added with the freedom of social media users to express their opinions, has influenced the spread of hate speech aimed at certain groups. Online based hate speech can be identified by the used of derogatory words in social media posts. Various studies on hate speech classification have been done, however, very few researches have been conducted on hate speech classification in the Indonesian language. This paper proposes a convolutional neural network method for classifying hate speech in tweets in the Indonesian language. Datasets for both the training and testing stages were collected from Twitter. The collected tweets were categorized into hate speech and non-hate speech. We used TF-IDF as the term weighting method for feature extraction. The most optimal training accuracy and validation accuracy gained were 90.85% and 88.34% at 45 epochs. For the testing stage, experiments were conducted with different amounts of testing data. The highest testing accuracy was 82.5%, achieved by the dataset with 50 tweets in each category.

show abstract

Section: Preprocessingmentioning

confidence: 99%

Hate Speech Classification in Indonesian Language Tweets by Using Convolutional Neural Network

Taradhita

Putra

2021

J. ICT Res. Appl.

View full text Add to dashboard Cite

show abstract

“…In this research, feature extraction is done by topic-based LDA method. LDA has some reasoning algorithms, one of which is Gibbs Sampling that have proven effective in conducting the topic sampling process [28]. In general, in the initialization process, Gibbs Sampling assigns the topic of each term randomly using a multinomial random function.…”

Section: Term Weigthing With Fuzzy Luhn's Gibbs Ldamentioning

confidence: 99%

“…LDA is a generative probabilistic model of a corpus, which documents are represented as random mixtures over latent topics, and each topic is characterized by a distribution over words [25]. A document in a corpus is not only identified as a single topic, but can be identified as several topics with their respective probabilities [26]- [28].…”

Section: Introductionmentioning

confidence: 99%

Complete agglomerative hierarchy document’s clustering based on fuzzy luhn’s gibbs latent dirichlet allocation

Prihatini¹,

Putra²,

Giriantari³

et al. 2019

IJECE

View full text Add to dashboard Cite

Agglomerative hierarchical is a bottom up clustering method, where the distances between documents can be retrieved by extracting feature values using a topic-based latent dirichlet allocation method. To reduce the number of features, term selection can be done using Luhn's Idea. Those methods can be used to build the better clusters for document. But, there is less research discusses it. Therefore, in this research, the term weighting calculation uses Luhn's Idea to select the terms by defining upper and lower cut-off, and then extracts the feature of terms using gibbs sampling latent dirichlet allocation combined with term frequency and fuzzy Sugeno method.The feature values used to be the distance between documents, and clustered with single, complete and average link algorithm. The evaluations show the feature extraction with and without lower cut-off have less difference. But, the topic determination of each term based on term frequency and fuzzy Sugeno method is better than Tsukamoto method in finding more relevant documents. The used of lower cut-off and fuzzy Sugeno gibbs latent dirichlet allocation for complete agglomerative hierarchical clustering have consistent metric values. This clustering method suggested as a better method in clustering documents that is more relevant to its gold standard.

show abstract

“…Latent Dirichlet Allocation (LDA) is a topic-based feature extraction method at term, document and corpus level [12]. Gibbs Sampling is one of the reasoning methods used for LDA [9]. LDA requires a sampling process that is carried out repeatedly until it reaches convergent conditions.…”

Section: Fuzzy Gibbs Latent Dirichlet Allocationmentioning

confidence: 99%

“…Wang, Zhou, Jin, Liu, and Lu used four methods: One-Hot Encoding, Term Frequency-Invers Document Frequency (TF-IDF) Weighting, word2vec and paragraph2vec which were applied to the short text classification with Naive Bayes (NB), Logistic Regression, Support Vector Machine (SVM), K-Nearest Neighbor (KNN) and Decision Tree as classifier [4] [7]. Prihatini, Putra, Giriantari, and Sudarma used TF-IDF as feature selection method and Fuzzy Gibbs Latent Dirichlet Allocation as feature extraction method for clustering news digital text, and resulted that the topic model gives better results in performing feature extraction than the classical model because the topic model distributes the term into all topics with different probabilities, while the classical model distributes the term only to one topic [8] [9]. This paper discusses three main factors in improving search engine performance.…”

Section: Introductionmentioning

confidence: 99%

Document Retrieval System Based on Topic Clustering Method

Prihatini¹,

Putra²,

Giriantari³

et al. 2018

Proceedings of the International Conference on Science and Technology (ICST 2018)

View full text Add to dashboard Cite

Document retrieval aims to find documents in a collection of unstructured text to meet the needs of user information. The search engine was required in the document retrieval system to perform the entire process automatically, starting from the processing of document text in the collection, feature selection, feature extraction, query text processing and search documents relevant to the query. There were three main factors in improving search engine performance: the feature selection method, the method of weighting features in document collections and the method of searching documents in the collection. In this paper, there were some methods used to improve the performance of search engines. For feature selection, Term Frequency-Invers Document Frequency based on Luhn's Idea was used for document features selection. For weighting features, Fuzzy Gibbs Latent Dirichlet Allocation was used for feature extraction method to weight the document features. To search documents that were relevant to the query, this paper used a Document Retrieval based on Topic Clustering method. Through this method, all documents were clustered based on feature weight obtained through feature extraction methods. Clusters that relevant to the query term combinations were selected and all documents in the cluster were displayed as search results. The result showed this method can retrieve set of documents in the cluster that relevant to the query. Therefore, this method could eliminate the query-document distance calculation function in the retrieval process, so it was hoped that the search process would run faster.

show abstract

Indonesian text feature extraction using gibbs sampling and mean variational inference latent dirichlet allocation

Cited by 10 publications

References 12 publications

Hate Speech Classification in Indonesian Language Tweets by Using Convolutional Neural Network

Hate Speech Classification in Indonesian Language Tweets by Using Convolutional Neural Network

Complete agglomerative hierarchy document’s clustering based on fuzzy luhn’s gibbs latent dirichlet allocation

Document Retrieval System Based on Topic Clustering Method

Contact Info

Product

Resources

About