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.

/pdf/hate-speech-classification-in-indonesian-language-tweets-by-4093zutxcx.pdf

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

In the learning management system, there are reviews from students of the learning process that has been done in a period. In this case, we use the review dataset to conduct sentiment analysis. The challenge of this dataset is the number of words that contain abbreviations and are not standard. So it challenges us to test the level of accuracy in the sentiment analysis process using several classification methods and sastrawi stemmer. Sastrawi stemmer is used to reduce features without changing the meaning data, Basic function of sastrawi is change words in the basic and eliminate nonessential or non-standard words with filtering concept. In the classification process, we use the SVM-PSO algorithm and compare it with other popular classification methods such as SVM, Naive Bayes and KNN. SVM-PSO is a combination of algorithms that is good to handle data with large dimensions and binary classification types. This is our reason for using SVM-PSO as the main classifer. Experimental results show that the use of sastrawi stemmer can reduce features by 32.58%. The accuracy of the classification process using SVM-PSO of 82.27% (with sastrawi stemmer) and 82.09% (without sastrawi stemmer), these results indicate that sastrawi stemmer influences the results of classification. SVM-PSO classification method has the highest level of accuracy compared to other classification methods, namely Naive Bayes gets an accuracy of 69.73%, K-NN gets an accuracy of 77.67% and SVM gets an accuracy of 81.52%. Based on the experimental results, SVM-PSO method has the best accuracy than any other method, and Sastrawi stemmer influences the level of accuracy.

Sentiment Analysis of Student Review in Learning Management System Based on Sastrawi Stemmer and SVM-PSO

Citizen science has emerged in many countries to contribute to the prompt resolution of individual field problems and has been shifted toward Information System (IS) research. In the domain of IS, a citizen report mechanism has been introduced in many local governments to understand regional problems based on the public participation. The rising of social media enforces many organizations including the local governments to utilize any information from the citizens including texts. Text mining has been utilized in various types of analyses such as sentiment analysis. However, it shows many challenges when it comes to the local context. The local context of words could cause various conflation errors that highly affect the learning task such as classification methods. This study aims to propose a context-based text processing and feed the proposed approach into a machine learning framework to classify the data of citizen reports-. The context-based text preprocessing utilized statistical- and semantic-based measurements to extract the local context and elaborate domain expertise to verify the misinterpretation for further text processing such as feature extractions. Subsequently, the n-gram language models together with the Term Frequency and Inverse Document Frequency schemes were performed to build the features. The result showed that the context-based text preprocessing improved the classification performance in majority classifiers in about 3% with the combinations of n-gram features. 

Contextual Text Analytics Framework for Citizen Report Classification: A Case Study Using the Indonesian Language

This research is about the comparison between the performances of single and multi-processing system. We will look that how jobs are processed in a uniprocessor and multiprocessor. We will work on process scheduling in both uni and multiprocessing systems. Some advantages and disadvantages are discussed here. Various workloads are tested to show what factors affecting their performance. Significant effort is made to make a fair comparison between them. The performance of various architectures is tested for uniprocessor system. Also much architectureis tested for the performance of multi-processor system.

UNI Processor and Multi-processor Performance Comparison

This research provides an analysis of tourism potential in Pekalongan Regency based on Twitter media so that it can provide input for related agencies to develop new potential tourism objects. Decision tree method with C4.5 is used to classify positive reviews where tourists will visit again and vice versa negative reviews with tweet data parameters related to tourism object, like location access, service satisfaction, conditions and functionality of existing facilities, and shopping experiences. The results of the review classification from the decision tree serve as input to the promotion strategy to be applied in the website media. Training data collection of 250 tweets related to the name of the tourism object in Pekalongan Regency was conducted and the experimental process was carried out in testing the model using RapidMiner 5.3. Stratified sampling with the C4.5 Decision Tree obtained the highest accuracy rate of 92% using fold = 6:4. 20 tweet sampling tests related to the Welo Asri Petungkriyono found 17 positive review tweets with the most words related to the condition and functionality of existing facilities and 3 negative reviews related to location access and service satisfaction.

Twitter Data as Decision Tree Parameter for Analysis of Tourism Potential Policies

Research in the field of Natural Language Processing (NLP) is currently increasing especially with the arrival of a new term that is "big data". The needs of the programming library that ready-touse becomes very important to speed up the phases of research. Some libraries that have already been mature is available but generally for English language and its dependently. So, it can't be used for other languages. Stemming is one of the basic processes that exist in NLP.Indonesian stemming algorithm that often used is ECS (Enhanced Confix Stripping). One of the libraries that already implemented the algorithm is Sastrawi. Results from the experiment show that the time of stemming processing by Sastrawi is still slow. Therefore, this research will optimize the speed of stemming processing using multiprocessing (MP). The data test are used in this research has manually taken form Wikipedia.The experiment results show that the MP technique can decrease the average time of stemming processing about 98.45%.

/pdf/multiprocessing-stemming-a-case-study-of-indonesian-stemming-41b92fgwsi.pdf

Rifkie Primartha

Papers

Multiprocessing Stemming: A Case Study of Indonesian Stemming