Publication Search

Fraud detection is used in various industries, including banking institutes, finance, insurance, government agencies, etc. Recent increases in the number of fraud attempts make fraud detection crucial for safeguarding financial information that is confidential or personal. Many types of fraud problems exist, including card-not-present fraud, fake Marchant, counterfeit checks, stolen credit cards, and others. An ensemble feature selection technique based on Recursive feature elimination (RFE), Information gain (IG), and Chi-Squared (X2) in concurrence with the Random Forest algorithm, was proposed to give research findings and results on fraud detection and prevention. The objective was to choose the essential features for training the model. The Receiver Operating Characteristic (ROC) Score, Accuracy, F1 Score, and Precision are used to evaluate the model's performance. The findings show that the model can differentiate between fraudulent transactions and those that are not, with an ROC Score of 95.83% and an Accuracy of 99.6%. The F1 Score of 99.6%% and precision of 100% further sustain the model's ability to detect fraudulent transactions with the least false positives correctly. The ensemble feature selection technique reduced training time and did not compromise the model's performance, making it a valuable tool for businesses in preventing fraudulent transactions.

Parkinson's disease (PD) is a neurodegenerative disorder causing a decline in dopamine levels, impacting the peripheral nervous system and motor functions. Current detection methods often identify PD at advanced stages. This study addresses early-stage detection using handwriting analysis, specifically exploring the PaHaW dataset for pen pressure and stroke movement data. Evaluating online and offline features, the research employs pre-trained CNN models (VGG 19 and AlexNet) for offline datasets, achieving an overall accuracy of 0.53. For online datasets, velocity, and acceleration features are extracted and classified using Gated Recurrent Unit (GRU), Long Short-Term Memory (LSTM), and recurrent neural networks (RNN), with GRU yielding the highest accuracy at 0.57. Notably, the convolution-based model C-Bi-GRU surpasses other architectures with a remarkable 0.75 accuracy, emphasizing its efficacy in early PD detection. These findings underscore the potential of handwriting analysis as a diagnostic tool for PD, contributing valuable insights for further research and development in medical diagnostics.

Dalam mengatasi masalah rendahnya kesadaran masyarakat dalam pemeriksaan kesehatan gigi dan mulut, machine learning dapat menjadi solusi praktis untuk mengatasi masalah tersebut. Solusi yang dapat diberikan berupa aplikasi yang dapat mengklasifikasikan dan mendeteksi secara dini gambaran penyakit gigi dan mulut. Ada banyak jenis pendekatan yang dapat digunakan untuk melakukan klasifikasi dan deteksi, namun yang paling banyak diteliti dan diterapkan adalah metode convolutional neural network yang merupakan salah satu dari beberapa jenis metode dari algoritma deep learning. Komputasi tanpa server dapat dilakukan dalam lingkungan yang terisolasi, ini dapat digunakan untuk menguji beberapa hipotesis secara paralel yang menguntungkan bagi pengembang, tetapi tantangan tetap ada pada komputasi tanpa server seperti jangka waktu dan kapasitas memori yang perlu ditangani, tetapi untuk saat ini komputasi tanpa server tampaknya telah menjadi alternatif yang layak untuk model proses pelatihan.

The advent of the Internet as an effective means for resource sharing has consequently, led to proliferation of adversaries, with unauthorized access to network resources. Adversaries achieved fraudulent activities via carefully crafted attacks of large magnitude targeted at personal gains and rewards. With the cost of over $1.3Trillion lost globally to financial crimes and the rise in such fraudulent activities vis the use of credit-cards, financial institutions and major stakeholders must begin to explore and exploit better and improved means to secure client data and funds. Banks and financial services must harness the creative mode rendered by machine learning schemes to help effectively manage such fraud attacks and threats. We propose HyGAMoNNE – a hybrid modular genetic algorithm trained neural network ensemble to detect fraud activities. The hybrid, equipped with knowledge to altruistically detect fraud on credit card transactions. Results show that the hybrid effectively differentiates, the benign class attacks/threats from genuine credit card transaction(s) with model accuracy of 92%.

Diabetes Mellitus is a hazardous disease, and according to the World Health Organization (WHO), diabetes will be one of the main causes of death by 2030. One of the most popular diabetes datasets is PIMA Indians, and this dataset has been widely tested on various machine learning (ML) methods, even deep learning (DL). But on average, ML methods are not able to produce good accuracy. The quality of the dataset and features is the most influential thing in this case, so deeper investment is needed to examine this dataset. This research will analyze and compare the PIMA Indians and Abelvikas datasets using the Random Forest (RF) method. The two datasets are imbalanced, in fact, the Abelvikas dataset is more imbalanced and has a larger number of classes so it is be more complex. The RF was chosen because it is one of the ML methods that has the best results on various diabetes datasets. Based on the test results, very contrasting results were obtained on the two datasets. Abelvikas had accuracy, precision, and recall, reaching 100%, and PIMA Indians only achieved 75% for accuracy, 87% for precision, and 80% for the best recall. Testing was done with 3, 5, 7, 10, and 15 tree number parameters. Apart from that, it was also tested with k-fold validation to get valid results. This determines that the features in the Abelvikas dataset are much better because more complete glucose features support them.

Skin is the largest organ in humans, it functions as the outermost protector of the organs inside. Therefore, the skin is often attacked by various diseases, especially cancer. Skin cancer is divided into two, namely benign and malignant. Malignant has the potential to spread and increase the risk of death. Skin cancer detection traditionally involves time-consuming laboratory tests to determine malignancy or benignity. Therefore, there is a demand for computer-assisted diagnosis through image analysis to expedite disease identification and classification. This study proposes to use the K-nearest neighbor (KNN) classifier and Gray Level Co-occurrence Matrix (GLCM) to classify these two types of skin cancer. Apart from that, the average filter is also used for preprocessing. The analysis was carried out comprehensively by carrying out 480 experiments on the ISIC dataset. Dataset variations were also carried out using random sampling techniques to test on smaller datasets, where experiments were carried out on 3297, 1649, 825, and 210 images. Several KNN parameters, namely the number of neighbors (k)=1 and distance (d)=1 to 3 were tested at angles 0, 45, 90, and 135. Maximum accuracy results were 79.24%, 79.39%, 83.63%, and 100% for respectively 3297, 1649, 825, and 210. These findings show that the KNN method is more effective in working on smaller datasets, besides that the use of the average filter also has a significant contribution in increasing the accuracy.

Phishing is a crime that uses social engineering techniques, both in deceptive statements and technically, to steal consumers' personal identification data and financial account credentials. With the new Phishing machine learning approach, websites can be recognized in real-time. K-Nearest Neighbor(KNN) and Naïve Bayes (NB) are popular machine learning approaches. KNN and NB have their own strengths and weaknesses. By combining the two, deficiencies can be covered. So this study proposes to combine K-Nearest Neighbor with Naïve Bayes to classify phishing websites. Based on the results of the accuracy test of the combination of KNN with k=8 and Naïve Bayes, a maximum accuracy of 93.44% is produced. This result is 6.25% superior compared to using only one classifier.

As software technology becomes more complex, software maintenance costs become more expensive. In connection with this, the development of software engineering makes the software system has many Composition choices that can be adjusted to the needs of the user. Error fixing involves analyzing Error Summary and modifying code. If bug-fixing steps are made as efficiently and effectively as possible then maintenance costs can be minimal. The purpose of this research is to establish a tool of machine learning for identifying Composition Error Summary and to find out the types of special Composition choices that can be used to save costs, time, and effort. In this study, the T-test was applied to appraise the analytical implication of conduct metrics when the “F-test” was taken to the Variance’s test. Classifiers used in this study are “All words” or “AW”, “Highly Informative Words” or “H-IW”, and “Highly Informative Words plus Bigram” or “H-WB”. Identical validation and Vexed validation techniques were used to calculate the effectiveness of machine learning tools. The results of this research denote that the instrument is competent for definitive Composition Error Summary and other Composition choices for definite Error Summary. This research determines the practicality of machine learning techniques in corrective issues relevant to Error summary. The result of this study also explained that Composition/non-Composition Error Summaries have contrasting aspects that can be accomplished by machine learning devices. The advanced tool could be upgraded in some areas to create it more powerful. The array identification section of the current study has limitations, an array with different words and Composition recognition tools tend to prefer Compositions with more words, so improvements to this could implicate consideration of the semantics of Error Summary, equivalent, and use of n-grams. Also, in using the technology of machine learning and Natural Language processing some advancements to be made to the present characterization structure so for future research it is highly recommended to clear up the first’s Error Summary before operating several operations in the present study.Composition Error Summary  

The Halal Product Assurance Agency (BPJPH) is an agency under the auspices of the Ministry of Religion with the task of ensuring the halalness of products in Indonesia. BPJPH has become a public concern after establishing the new halal logo. On February 10, 2022 the new halal logo was ratified by the Head of BPJPH, Muhammad Aqil Irham. This has become a topic of public discussion either directly or through social media, one of which is social media twitter. The number of opinion tweets about the change of the halal logo can be used as a data source to obtain information about public opinion on the change of the halal logo through sentiment analysis. Sentiment analysis can be done by machine learning approach, one of these is the SVM algorithm . In this research, oversampling and undersampling are applied to handle data that has an unbalanced sentiment class. The results showed that the Support Vector Machine (SVM) model using oversampling training data got the highest accuracy, recall, precision, and f1-score, namely 71% accuracy, 67% precision, 61% recall, and 61% f1-score while training using undersampling training data has the lowest performance, namely getting 56% accuracy, 51% precision, 57% recall, and 52% f1-score.

Tujuan Utama: Tujuan dari penelitian ini adalah untuk mengembangkan algoritma deteksi anomali yang lebih efektif dan akurat menggunakan Extended Isolation Forest (EIF) dan mengimplementasikannya ke dalam platform sumber terbuka Machine Learning (ML) H2O-3. Background problem: Algoritma Isolation Forest (IF) asli menghadirkan bentuk deteksi baru, meskipun algoritme mengalami bias yang berasal dari percabangan pohon. Perpanjangan algoritme menghilangkan bias dengan menyesuaikan percabangan, dan algoritme asli hanya menjadi kasus khusus. EIF diimplementasikan ke dalam platform sumber terbuka ML H2O-3. kebaharuan: Kebaruan dari penelitian ini adalah penggunaan algoritma EIF dalam deteksi anomali. Selain itu, penelitian ini juga mengimplementasikan EIF ke dalam platform sumber terbuka ML H2O-3 untuk dijalankan pada sistem komputasi terdistribusi dengan pustaka Map/Reduce. Research Method: Penelitian ini menggunakan metode deteksi anomali dengan fokus pada algoritma EIF.  temuan: Hasil pengujian menunjukkan bahwa Extended Isolation Model perlu disesuaikan. Tes kinerja deteksi anomali mengungkapkan sedikit ketidaksempurnaan dalam deteksi struktur data jika dibandingkan dengan satu-satunya implementasi algoritma Python yang tersedia. Hasil ujian untuk tahap evaluasi dinyatakan lulus dan waktu komputasi secara logaritmik lebih kecil dengan jumlah utas.  Kesimpulan: pada penelitian selanjutnya, algoritma dapat ditingkatkan lebih lanjut dengan menskalakan anomali deteksi untuk data dimensi tinggi. Ini dapat diimplementasikan dengan menambahkan parameter lain yang memungkinkan metode pemilihan fitur dalam perhitungan..

Main Objective: Tujuan dari penelitian ini adalah untuk mengetahui penyebab dari flaky test yang paling umum terjadi pada proyek Python dengan membandingkan penelitian yang berfokus pada proyek Java sebelumnya. Background problem: Tes tidak stabil dapat gagal atau lulus tanpa ada perubahan pada kode yang diuji, ini dapat menghancurkan kepercayaan pengembang pada rangkaian pengujian dan jika diabaikan menyebabkan bug dalam kode yang dirilis.  Novelty: penelitian menggunakan pendekatan empiris dari proyek Python open-source paling populer di GitHub. Sejumlah 197 komitmen dengan kata kunci yang menunjukkan kelemahan uji diperiksa secara manual dan dikategorikan menurut akar penyebab kelemahan masing-masing. Research Method: metode analisis pengujian flakiness digunakan dengan urutan tahap filtering commit dan dua analisis. Finding/Result: hasil penelitian dibandingkan dengan studi proyek Java sebelumnya, dan ditemukan dua penyebab kelemahan yakni presisi dan pelatihan (jaringan Machine learning). Flakiness presisi disebabkan oleh pernyataan dengan ambang batas yang terlalu tinggi atau terlalu rendah. Kelemahan pelatihan disebabkan oleh pengaturan pelatihan yang salah dari jaringan Machine learning dalam pengujian. Sebagian besar tes dalam proyek Python ditemukan tidak stabil karena masalah dengan menunggu asinkron, presisi, dan jaringan.  Conclusion: Developer Python dimasa depan akan mendapat manfaat dari pengetahuan tentang jebakan umum yang dapat menyebabkan kelemahan dalam rangkaian pengujian mereka. Hasil dari penelitian ini dapat digunakan sebagai referensi bagi para peneliti di masa depan dengan area penlitian tes flakiness atau area serupa lainnya..

Property price determination is a complex challenge influenced by various factors, thus requiring an effective method for accurate prediction to support investment decision-making. In the current digital era, conventional approaches are being replaced by data-driven and artificial intelligence methods, where Linear Regression remains a popular choice due to its simplicity and effectiveness in modeling linear relationships. This study aims to analyze the relationship between the physical characteristics of a house and its selling price, and to build an accurate predictive model using the Linear Regression algorithm. A quantitative method was used, focusing on Building Area , Number of Rooms, and Building Age  against the House Selling Price. Correlation analysis results show that Building Area has the strongest correlation (0.81) with price, while Building Age shows a negative correlation (-0.52). The Linear Regression model demonstrated very strong and stable performance. The model achieved an R² Score of 0.9396 on the testing data, meaning 93.96% of house price variability can be explained by the model. Furthermore, the low MAE of only 11.31 million rupiah indicates a small prediction error, and the consistency of R² scores confirms that the model does not suffer from overfitting. This study concludes that the Linear Regression model provides excellent, stable, and reliable prediction performance for projecting house selling prices

DDOS (Distribute Denial of Service) is a type of structured attack. This attack has been around since 1990. DDoS attacks are capable of paralyzing servers by flooding network traffic and causing it to go down. To overcome this problem, the way to detect DDoS attacks has several methods and algorithms, one of which is the Artificial Neural Network algorithm and uses the Machine learning method due to the fast computing process, high accuracy, and this research uses the SKKNI research method Number 299 of 2020. The analysis was carried out uses training data from the latest dataset, namely CICIDS2017, which is a development of a previously existing dataset. DDoS attack detection testing using the confusion matrix method obtained bot precision of 0.99, recall of 0.99, and f1score of 0.99, 3

Pengelolaan sampah yang buruk telah menjadi permasalahan yang masih dialami  di seluruh dunia, tak terkecuali Indonesia. Pengedukasian mengenai pengelolaan sampah perlu ditingkatkan terutama pada anak-anak dengan memanfaatkan perkembangan teknologi yang ada seperti teknologi smartphone. Berdasarkan masalah tersebut, dikembangkanlah aplikasi edukasi mengenai pengelolaan sampah yang ditujukan untuk anak sekolah dasar dengan memanfaatkan teknologi Android dan machine learning. Dengan tingkat akurasi sebesar 90% pada training dan testing dataset, pemanfaatan teknologi machine learning ini akan efektif untuk membantu anak-anak dalam mendeteksi sampah yang ditemukan ketika bereksplorasi. penelitian ini menggunakan metode Waterfall yang terdiri dari 5 tahapan yaitu requirement,  design, implementation, testing dan maintenance. Dengan menggunakan metode pengujian Black Box dan Beta Test, hasil pengujian yang didapatkan pada pengujian Black Box telah memenuhi semua skenario yang ada, dan melalui Beta Test mendapatkan respons positif serta beberapa saran dari calon pengguna untuk pengembangan aplikasi kedepannya. Dengan hasil yang didapatkan, aplikasi edukasi mengenai pengelolaan sampah dapat menjadi sarana untuk membantu anak-anak dalam memahami betapa pentingnya pengelolaan sampah melalui media pembelajaran interaktif.