Algoritma LSTM-CNN untuk Binary Klasifikasi dengan Word2vec pada Media Online
Abstract
Media online banyak menghasilkan berbagai macam berita, baik ekonomi, politik, kesehatan, olahraga atau ilmu pengetahuan. Di antara itu semua, ekonomi adalah salah satu topik menarik untuk dibahas. Ekonomi memiliki dampak langsung kepada warga negara, perusahaan, bahkan pasar tradisional tergantung pada kondisi ekonomi di suatu negara. Sentimen yang terkandung dalam berita dapat mempengaruhi pandangan masyarakat terhadap suatu hal atau kebijakan pemerintah. Topik ekonomi adalah bahasan yang menarik untuk dilakukan penelitian karena memiliki dampak langsung kepada masyarakat Indonesia. Namun, masih sedikit penelitian yang menerapkan metode deep learning yaitu Long Short-Term Memory dan CNN untuk analisis sentimen pada artikel finance di Indonesia. Penelitian ini bertujuan untuk melakukan pengklasifikasian judul berita berbahasa Indonesia berdasarkan sentimen positif, negatif dengan menggunakan metode LSTM, LSTM-CNN, CNN-LSTM. Dataset yang digunakan adalah data judul artikel berbahasa Indonesia yang diambil dari situs Detik Finance. Berdasarkan hasil pengujian memperlihatkan bahwa metode LSTM, LSTM-CNN, CNN-LSTM memiliki hasil akurasi sebesar, 62%, 65% dan 74%.
Kata Kunci — LSTM, sentiment analysis, CNN
Online media produce a lot of various kinds of news, be it economics, politics, health, sports or science. Among them, economics is one interesting topic to discuss. The economy has a direct impact on citizens, companies, and even traditional markets depending on the economic conditions in a country. The sentiment contained in the news can influence people's views on a matter or government policy. The topic of economics is an interesting topic for research because it has a direct impact on Indonesian society. However, there are still few studies that apply deep learning methods, namely Long Short-Term Memory and CNN for sentiment analysis on finance articles in Indonesia. This study aims to classify Indonesian news headlines based on positive and negative sentiments using the LSTM, LSTM-CNN, CNN-LSTM methods. The dataset used is data on Indonesian language article titles taken from the Detik Finance website. Based on the test results, it shows that the LSTM, LSTM-CNN, CNN-LSTM methods have an accuracy of, 62%, 65% and 74%.
Keywords — LSTM, sentiment analysis, CNN
Full Text:
PDFReferences
Detiknet.com, 2020, Riset: Ada 175,2 Juta Pengguna Internet di Indonesia, https://inet.detik.com/cyberlife/d-4907674/riset-ada-1752-juta-pengguna-internet-di-indonesia.
Meng, X., Liu, M., Wu, Q., 2020, Prediction of Rice Yield via Stacked LSTM, International Journal of Agricultural and Environmental Information Systems (IJAEIS), No. 1, Vol. 11, Hal. 86–95
Reis, J., Benevenuto, F., Olmo, P., Prates, R., Kwak, H., An, J., 2015, Breaking the news: First impressions matter on online news, The 9th International Conference on Web and Social Media, oxford, 26-29 Mei
Zamahsyari., Nurwidyantoro, A., 2017, Sentiment analysis of economic news in Bahasa Indonesia using majority vote classifier, 2016 International Conference on Data and Software Engineering (ICoDSE 2016), Denpasarr, 26 - 27 Oktober
Joshi, K., N Bharathi, H., Rao, J., 2016, Stock Trend Prediction Using News Sentiment Analysis, International Journal of Computer Science & Information Technology, No. 3, Vol. 8, Hal. 1-8
Zubiaga, A., Kochkina, E., Liakata, M., Procter, R., Lukasik, M., 2016, Structure of Social Media Conversations, arXiv preprint, arXiv:1609.09028
Nugraheni, E., 2020, Indonesian Twitter Data Pre-processing for the Emotion Recognition, 2019 International Seminar on Research of Information Technology and Intelligent Systems (ISRITI), Yogyakarta, 5 -6 Desember
Kochkina, E., Liakata, M., Augenstein, I., 2017, Turing at SemEval-2017 Task 8: Sequential Approach to Rumour Stance Classification with Branch-LSTM, the 11th International Workshop on Semantic Evaluation (SemEval-2017), Vancuver, 3-4 Agustus
Augenstein, I., Rockt, T., Vlachos, A., Bontcheva, K., 2016, Stance Detection with Bidirectional Conditional Encoding, arXiv preprint, arXiv:1606.05464
Jelodar, H., Wang, Y., Orji, R., Huang, H., 2020, Deep Sentiment Classification and Topic Discovery on Novel Coronavirus or COVID-19 Online Discussions: NLP Using LSTM Recurrent Neural Network Approach, IEEE Journal of Biomedical and Health Informatics, doi: 10.1109/JBHI.2020.3001216
Murthy, G. S. N., Allu, S. R., Andhavarapu, B., Bagadi, M., Belusont, M., 2020, Text based Sentiment Analysis using LSTM, IJERT, No. 5, Vol. 9, Hal. 299–303
Jain, D., Kumar, A., Garg, G., 2020, Sarcasm detection in mash-up language using soft-attention based bi-directional LSTM and feature-rich CNN, Applied Soft Computiing Journal, Vol. 91, Hal. 106198, 2020, doi: 10.1016/j.asoc.2020.106198.
Kane, S. N., Mishra, A., Dutta, A. K., 2016, Preface: International Conference on Recent Trends in Physics (ICRTP 2016), International Conference on Recent Trends in Physics 2016 (ICRTP2016), Inore - India, 13-14 Febrari
Nurrohmat, M. A., A. Sn, 2019, Sentiment Analysis of Novel Review Using Long Short-Term Memory Method, IJCCS (Indonesian Journal of Computing and Cybernetics Systems, No. 3, Vol. 13, Hal. 209–218
Kurniasari, L., Setyanto, A., 2020, Sentiment Analysis using Recurrent Neural Network, Journal of Physics: Conference Series, Vol. 1471, Bukit Tinggi, 18 Oktober, doi: 10.1088/1742-6596/1471/1/012018.
Manaswi, N. K., 2018, Deep Learning with Applications Using Python, Apress Media, New York
Sosa, P. M., 2017, Twitter Sentiment Analysis using combined LSTM-CNN Models, Academia.edu, Hal. 1–9
DOI: https://doi.org/10.24076/citec.2021v8i1.264
Refbacks
- There are currently no refbacks.
Indexed by:
Dedicated to:
