Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Muhammad Habibi has a total of 44 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    User Requirement Recommendation Model for Waste Reporting Platforms Based on UX Topics and Sentiment Analysis

    by Irmma Dwijayanti, Alfirna Rizqi Lahitani, Muhammad Habibi

    Waste management remains a critical issue in Indonesia, as emphasized in the RPJMN 2025–2029. Ineffective collection and processing services, coupled with limited public participation, continue to hinder progress. Meanwhile, social media has emerged as a primary channel for citizens to express complaints and reports on waste, yet the unstructured nature of comments poses challenges for integration into official reporting systems. This study proposes a user requirements recommendation model based on social media data by integrating sentiment analysis, topic modeling, and rule-based recommendation. Data were collected from YouTube and TikTok comments. Sentiment classification was performed using Support Vector Machine (SVM), while Latent Dirichlet Allocation (LDA) was employed for topic modeling, with results mapped onto the UX Honeycomb dimensions. Recommendation rules were then formulated by combining sentiment polarity with UX dimensions. The SVM model achieved an average accuracy of 87.5% with balanced precision, recall, and F1-score. LDA produced 15 coherent topics, which were distributed across seven UX dimensions. The integration revealed that the main user requirements include transparency in report follow-up through real-time notifications and clear status updates. Additional recommendations involve simplifying the reporting process, providing auto-fill features, improving visual design, and establishing a user appreciation system. The findings demonstrate the potential of leveraging social media comments to systematically capture user requirements, offering practical insights for developers to design waste reporting platforms that are effective, user-friendly, and responsive to community expectations.

  • Open Access

    Mapping User Dissatisfaction in Mobile Banking Applications Using Ensemble Clustering LDA and LSA

    by Kartikadyota Kusumaningtyas, Alfirna Rizqi Lahitani, Irmma Dwijayanti, Muhammad Habibi

    Mobile banking has become one of the most popular choices compared to other online banking services. Google Play Store is an online application store platform provides a review section for users to give ratings and comments on the applications they use. Positive reviews typically contain good experiences that reflect user satisfaction, while negative reviews usually contain poor experiences that indicate complaints and user dissatisfaction. However, Google Play Store does not yet have a feature to automatically map the main topics in both positive and negative reviews. Specifically for negative reviews, this can make it difficult for developers to understand the root problems and take appropriate corrective actions. In some situations, negative reviews need to be handled more quickly. Slow handling of negative reviews can impact the decline in reputation and customer loyalty. This research aims to identify user dissatisfaction topics based on negative reviews of several popular mobile banking applications in Indonesia, namely BCA Mobile and BRImo.


  • Open Access

    Predictive Modeling of Air Quality Levels Using Decision Tree Classification: Insights from Environmental and Demographic Factors

    by I Gede Iwan Sudipa, Muhammad Habibi, Ery Setiyawa Jullev Atmadji, Ika Arfiani

    Air pollution poses a significant global challenge, adversely impacting public health and environmental sustainability. Understanding the factors influencing air quality is essential for developing effective mitigation strategies. This study aims to analyse key environmental and demographic factors, such as PM2.5 concentration, population density, and proximity to industrial areas, to predict air quality levels using a Decision Tree model. The dataset, comprising 5000 samples, was pre-processed by encoding the target variable and applying Z-score normalization to numerical features. The model was trained on 80% of the data and evaluated on the remaining 20%, achieving an accuracy of 93%. Evaluation metrics, including a classification report and confusion matrix, demonstrated the model's effectiveness in distinguishing between four air quality categories: Good, Moderate, Poor, and Hazardous. PM2.5 emerged as the most critical predictor, followed by demographic and industrial factors. These findings underscore the potential of machine learning models in providing actionable insights for air quality management. The results contribute to public policy by highlighting the need for targeted interventions in high-risk areas and the importance of incorporating environmental data into urban planning. Future work should focus on expanding the feature set and exploring ensemble techniques to further enhance predictive accuracy and robustness.


  • Open Access

    Analisis Tren Topik dalam Ulasan Negatif Aplikasi M-Banking Menggunakan Latent Dirichlet Allocation

    by Kartikadyota Kusumaningtyas, Irmma Dwijayanti, Alfirna Rizqi Lahitani, Muhammad Habibi

    Mobile banking atau M-banking menjadi semakin populer seiring dengan meluasnya penggunaan ponsel pintar. Pertumbuhan ini didorong oleh beberapa faktor, seperti kebijakan pemerintah melalui Gerakan Nasional Non-Tunai (GNTT) dan inovasi dari bank. Latar belakang penelitian ini berangkat dari pentingnya merespons keluhan pengguna terhadap aplikasi M-banking. Ulasan negatif mencerminkan masalah yang dialami pengguna dan bisa memengaruhi kepercayaan terhadap layanan. Sayangnya, platform seperti Google Play Store tidak menyediakan fitur untuk mengidentifikasi tren dari ulasan negatif. Oleh karena itu, penelitian ini menggunakan metode Latent Dirichlet Allocation (LDA) untuk memodelkan tren topik dalam ulasan negatif guna memberikan wawasan bagi penyedia layanan untuk meningkatkan kualitas aplikasi mereka. Penelitian ini dilakukan melalui beberapa tahap, dimulai dengan pengumpulan data ulasan negatif dari tiga aplikasi M-banking populer. Selanjutnya data akan melalui tahap preprocessing, meliputi: tokenizing, stopwords removal, dan stemming. Sentimen dari ulasan dianalisis menggunakan algoritma Support Vector Machine (SVM) dengan akurasi mencapai 93%, untuk memisahkan ulasan positif dan negatif. Selanjutnya, LDA digunakan untuk memodelkan topik pada ulasan negatif, dengan mengidentifikasi sejumlah topik optimal melalui Coherence Score, yang menunjukkan struktur topik yang logis dan terorganisir. Hasil penelitian menunjukkan bahwa pada BRImo, topik yang dominan adalah biaya dan kecepatan layanan aplikasi. Pada BCA mobile, pengguna lebih banyak membahas fitur dan kemudahan penggunaan aplikasi, sedangkan pada Livin’ by Mandiri, topik utama yang dibahas berkaitan dengan fitur transfer dan jam transaksi. Kesimpulan dari penelitian ini adalah bahwa metode LDA berhasil digunakan untuk menemukan tren utama dari ulasan negatif pengguna, yang diharapkan dapat membantu bank dalam meningkatkan kualitas layanan dan keamanan aplikasi mobile banking.


  • Open Access

    Ekstraksi Aspek Aksesibilitas untuk Peningkatan Pengalaman Pengguna Menggunakan NER dengan CNN dan LSTM

    by Irmma Dwijayanti, Alfirna Rizqi Lahitani, Kartikadyota Kusumaningtyas, Muhammad Habibi

    Transportasi online memberikan dampak positif bagi sebagian besar masyarakat, namun penggunaan aplikasi oleh penyandang disabilitas masih menghadapi sejumlah tantangan. Belum adanya fasilitas yang memadai dan pengalaman pengguna yang baik menjadi kendala utama. Realitas ini menunjukkan bahwa perlunya perhatian khusus terhadap prinsip aksesibilitas untuk meningkatkan pengalaman pengguna dan kenyamanan bagi penyandang disabilitas. Melalui ulasan pengguna dapat diidentifikasi aspek-aspek aksesibilitas untuk mendukung peningkatan pengalaman pengguna. Penelitian ini bertujuan mengekstraksi informasi dari ulasan pengguna terkait aksesibilitas menggunakan metode NER dengan pendekatan CNN dan LSTM. Data yang dikumpulkan melalui web scraping terdiri dari 6.255 ulasan aplikasi Gojek, Grab, Maxim, dan Indriver. Hasil evaluasi menunjukkan bahwa kedua model memiliki akurasi tinggi yaitu CNN 99,84%, dan LSTM 99,48%. Namun memerlukan perbaikan dalam mendeteksi entitas yang jarang muncul atau berkonteks kompleks. Hasil analisis menunjukkan bahwa ulasan lebih banyak membahas fitur aplikasi dan keluhan yang berkaitan dengan aksesibilitas. CNN lebih efektif dalam menangkap pola spesifik, sedangkan LSTM lebih kuat dalam menangkap variasi kata.

  • Open Access

    ANALISIS PROYEKSI KEBUTUHAN TENAGA KERJA BERDASARKAN SKILLS YANG DIBUTUHKAN MENGGUNAKAN ALGORITMA NAIVE BAYES CLASSIFIER

    by Nur Azizah Firdausa, Ribka Rifanny Br Girsang, Dela Oktaviana, Astr Wahyuningsiam, Muhammad Habibi

    Pada Agustus 2023, Indonesia menghadapi jumlah pengangguran mencapai 7,86 juta orang walaupun tidak memungkiri bahwa persentase pengangguran telah menurun dari tahun sebelumnya. Dari data tersebut dikelompokkan menjadi empat kelompok yaitu pengangguran melibatkan mereka yang mencari pekerjaan, berusaha mendirikan usaha, merasa sulit mendapatkan pekerjaan, dan bahkan yang sudah bekerja namun belum mulai. Pandemi Covid-19 mengubah paradigma pekerjaan menjadi remote, namun kebutuhan akan informasi pekerjaan tetap menjadi kunci. Proyeksi kebutuhan tenaga kerja memberikan wawasan jangka panjang tentang sektor dan bidang yang menjanjikan, memandu pencari kerja untuk mengembangkan keterampilan sesuai tren pasar kerja. Penelitian ini dilakukan dengan menggunakan klasifikasi naive bayes, yaitu sebuah metode klasifikasi teks yang bergantung pada kemungkinan kata kunci untuk membandingkan dokumen pelatihan dan dokumen pengujian. Dengan metode klasifikasi ini diharapkan dapat membantu mengurangi tingkat pengangguran dan menyelaraskan keterampilan individu dengan kebutuhan industri, memberikan kontribusi pada kebijakan pendidikan dan pelatihan untuk membuat keputusan karir yang cerdas di era digital.

  • Open Access

    Metode Latent Dirichlet Allocation Untuk Menentukan Topik Pada Review Drama Korea

    by Alfun Roehatul Jannah, Ria Kristi, Muhammad Habibi

    Hallyu Wave, yang melibatkan penyebaran budaya dan media populer Korea Selatan, telah berkembang pesat selama dua dekade terakhir. Selain industri hiburan seperti K-pop dan K-drama, fenomena ini juga merambah ke sektor makanan dan K-beauty. Drama Korea, sebagai inti dari Hallyu, telah menjadi fenomena global dengan basis penggemar yang terus berkembang di seluruh dunia. Survei global tahun 2022 menunjukkan bahwa 36 persen responden di 26 negara menganggap drama Korea sangat populer di negara mereka. Di Indonesia, film dan drama Korea tetap menjadi favorit, dengan 72 persen penonton streaming memilihnya pada layanan OTT sepanjang 2022. Viu mendominasi sebagai platform streaming drakor paling populer dengan 57 persen penggunaan, diikuti oleh Netflix, Telegram, dan WeTv. Penelitian ini fokus pada analisis data review drama Korea dari 2015–2023 menggunakan metode Latent Dirichlet Allocation (LDA). Tujuannya adalah memberikan pemahaman mendalam tentang aspek kritis seperti akting, alur cerita, dan sinematografi. Dengan LDA, penelitian ini bertujuan mengidentifikasi topik-topik terkait elemen-elemen tersebut, memberikan wawasan spesifik tentang preferensi penonton. Dari hasil penelitian yang dilakukan, muncul 10 topik ideal dari 20 topik yang ada untuk memastikan konsistensi topik menggunakan topic coherence. Dari hasil topic coherence untuk 20 topik tersebut, dapat disimpulkan bahwa skor topik keseluruhan untuk topik 10 adalah 0,527 yang memberikan hasil topic modelling yang ideal dan sesuai dengan ketetapan.

  • Open Access

    Improving Mental Health Diagnostics through Advanced Algorithmic Models: A Case Study of Bipolar and Depressive Disorders

    by Adityo Permana Wibowo, Medi Taruk, ⁠⁠Thomas Edyson Tarigan, Muhammad Habibi

    Indonesian Journal of Data and Science


  • Open Access

    Tweet Analysis of Mental Illness Using K-Means Clustering and Support Vector Machine

    by Kartikadyota Kusumaningtyas, Muhammad Habibi, Irmma Dwijayanti, Retno Sumiyarini

    Purpose: Social media, particularly Twitter, provides a venue for individuals to share their thoughts. The public's perception of mental illnesses is often debated on Twitter. So yet, no evaluation of community tweets connected to data on mental health conditions has been performed. The purpose of this study is to examine tweets linked to mental illnesses in Indonesia in order to identify the themes of conversation and the polarity trends of these tweets.Design/methodology/approach: To address this issue, the K-Means Clustering algorithm is utilized to aggregate tweet data that is used to find themes of conversation. The emotion polarity value of each cluster result was then determined using the Support Vector Machine (SVM) approach.Findings/results: This study generated five topic clusters based on tweets about mental illness. While sentiment analysis revealed that all clusters had more negative sentiment classes than positive. Cluster 4 and Cluster 5 had the highest number of negative sentiment values. These clusters emphasize the necessity of consulting with psychiatrists and psychologists if people have mental health disorders, as well as financing for mental health disorder treatment through BPJS Kesehatan services.Originality/value/state of the art: The analysis was done in two stages: data grouping to find themes of conversation using K-Means clustering and SVM to look for positive and negative polarity values associated to twitter data about mental illness.


  • Customer Experience Analysis Skincare Products Through Social Media Data Using Topic Modeling and Sentiment Analysis

    by Muhammad Habibi, Kartikadyota Kusumaningtyas

    Currently, skin care products (skincare) are popular among the public. Both men and women are interested in buying skin care products. Moreover, there are many brands of skin care products that are divided into several types of facial and body care, such as moisturizers, toners, cleansers, and masks. Therefore, many consumers take the time to find information, for example, in terms of price, quality, and brand for decision-making. A lot of useful information is in the form of Twitter messages known as tweets which are sent from people who use skin care products because Twitter is one of the online social media where users can share their opinions and experiences. However, consumers still have to spend a lot of time searching, reading, and understanding the comprehensive collection of tweets before buying skin care products.The purpose of this study is to analyze customer experience, analyzing automated tweets about skin care products. Tweets about skin care products will be subjected to a topic modeling process to find out what topics are being discussed. In addition, the topics that have been obtained will be subject to sentiment analysis in the form of positive and negative messages for skin care products. Consumers who are app users don’t waste time reading and analyzing large amounts of data manually and they can decide to buy skin care products more easily.The results of this study obtained 14 topics of discussion related to skincare. Meanwhile, the sentiment analysis results of 14 topics resulted in more positive sentiment class tweets overall. It related the category topic that has the number of tweets to the importance of skincare. In addition, categories related to ingredients for skincare products from nature, namely fruits and spices, are the topics that have the second highest number of tweets. The results of the analysis of tweets related to user experience on Twitter, it was found that users prefer skincare products that use ingredients from nature.