Jurnal Transformatika
Not a member yet
319 research outputs found
Sort by
Social Assistance Recipient Decision Support System with AHP and MOORA
No full textPoverty is a multidimensional problem that requires prompt and appropriate handling to maintain a dignified human life. In Manyaran Sub-district, Semarang City, the distribution of social assistance often faces obstacles due to limited human resources and a manual selection process for recipients. Therefore, a Decision Support System (DSS) is needed to assist the selection process in a more objective and efficient manner. This study aims to develop a DSS for determining social assistance recipients in Manyaran Sub-district by combining the Analytic Hierarchy Process (AHP) and Multi-Objective Optimization on the basis of Ratio Analysis (MOORA) methods. AHP is utilized to determine the weight of each criterion, while MOORA is used to calculate the final score of each recipient candidate. The results show that among the ten analyzed candidates, the individual coded P09 achieved the highest final score of 0.575. The top five candidates with the highest scores were declared eligible to receive social assistance, while the others were declared ineligible. The application of the AHP and MOORA methods in this DSS effectively improves the accuracy, objectivity, and efficiency of the selection process for social assistance recipients in Manyaran Sub-district.
Poverty is a multidimensional problem that requires prompt and appropriate handling to maintain a dignified human life. In Manyaran Sub-district, Semarang City, the distribution of social assistance often faces obstacles due to limited human resources and a manual selection process for recipients. Therefore, a Decision Support System (DSS) is needed to assist the selection process in a more objective and efficient manner. This study aims to develop a DSS for determining social assistance recipients in Manyaran Sub-district by combining the Analytic Hierarchy Process (AHP) and Multi-Objective Optimization on the basis of Ratio Analysis (MOORA) methods. AHP is utilized to determine the weight of each criterion, while MOORA is used to calculate the final score of each recipient candidate. The results show that among the ten analyzed candidates, the individual coded P09 achieved the highest final score of 0.575. The top five candidates with the highest scores were declared eligible to receive social assistance, while the others were declared ineligible. The application of the AHP and MOORA methods in this DSS effectively improves the accuracy, objectivity, and efficiency of the selection process for social assistance recipients in Manyaran Sub-district
Boosting Performance Classification KNN Customer Loyalty with Chi-Square and Information Gain
Full text linkUnderstanding customer purchasing behavior is essential for predicting customer loyalty, which directly impacts a company\u27s long-term success. This research aims to determine the effect of chi-square and information gain feature selection in optimizing customer loyalty classification performance, compared to pure kNN. Using a public customer purchasing behavior dataset from Kaggle, containing 10,000 data, 12 attributes with loyalty_status as the label (Gold, Regular, Silver). Evaluating performance by accuracy, kappa, classification error, recall, precision, and RMSE. The highest accuracy 91.99% was obtained by kNN k=3 with information gain, kappa 0.844, precision 95.44%, recall 86.30%, with the lowest classification error 8.01% and the second lowest RMSE 0.245, after kNN k=3 with chi-square. Results show that feature selection has a positive impact on classification, increasing accuracy and reducing errors, with the combination of the kNN k=3 method and information gain proving successful in obtaining high accuracy in classifying customer loyalty.Understanding customer purchasing behavior is essential for predicting customer loyalty, which directly impacts a company\u27s long-term success. This research aims to determine the effect of chi-square and information gain feature selection in optimizing customer loyalty classification performance, compared to pure kNN. Using a public customer purchasing behavior dataset from Kaggle, containing 10,000 data, 12 attributes with loyalty_status as the label (Gold, Regular, Silver). Evaluating performance by accuracy, kappa, classification error, recall, precision, and RMSE. The highest accuracy 91.99% was obtained by kNN k=3 with information gain, kappa 0.844, precision 95.44%, recall 86.30%, with the lowest classification error 8.01% and the second lowest RMSE 0.245, after kNN k=3 with chi-square. Results show that feature selection has a positive impact on classification, increasing accuracy and reducing errors, with the combination of the kNN k=3 method and information gain proving successful in obtaining high accuracy in classifying customer loyalty.
KLASIFIKASI SAMPAH ORGANIK DAN NON ORGANIK MENGGUNAKAN TRANSFER LEARNING
Full text linkPengelolaan sampah di Indonesia menghadapi tantangan serius dengan 7,2 juta ton sampah belum terkelola dengan baik dari 202 kabupaten/kota, mencemari lingkungan dan menghambat daur ulang berkelanjutan. Pemilahan sampah organik dan anorganik yang masih dilakukan secara manual rentan terhadap kesalahan manusia dan tidak efisien. Penelitian ini mengembangkan model klasifikasi sampah organik dan anorganik menggunakan metode transfer learning dengan tiga arsitektur CNN: VGG16, MobileNetV2, dan ResNet50V2. Dataset diambil dari kaggle Waste Classification Data yang telah melalui proses preprocessing. Hasil eksperimen menunjukkan bahwa MobileNetV2 unggul dengan akurasi 90,13%, presisi 96,25%, dan F1-Score 87,88%, waktu inferensi 127,76 ms. Arsitektur ini memberikan keseimbangan optimal antara performa tinggi dan efisiensi komputasi, sehingga ideal diterapkan pada perangkat pintar seperti ponsel dan sistem IoT dalam konteks manajemen sampah perkotaan. Penelitian ini menegaskan efektivitas transfer learning dalam membangun sistem klasifikasi sampah yang cerdas dan efisien untuk mendukung program pemilahan sampah di tingkat rumah tangga dan institusi.
Pengelolaan sampah di Indonesia menghadapi tantangan serius dengan 7,2 juta ton sampah belum terkelola dengan baik dari 202 kabupaten/kota, mencemari lingkungan dan menghambat daur ulang berkelanjutan. Pemilahan sampah organik dan anorganik yang masih dilakukan secara manual rentan terhadap kesalahan manusia dan tidak efisien. Penelitian ini mengembangkan model klasifikasi sampah organik dan anorganik menggunakan metode transfer learning dengan tiga arsitektur CNN: VGG16, MobileNetV2, dan ResNet50V2. Dataset diambil dari kaggle Waste Classification Data yang telah melalui proses preprocessing. Hasil eksperimen menunjukkan bahwa MobileNetV2 unggul dengan akurasi 90,13%, presisi 96,25%, dan F1-Score 87,88%, waktu inferensi 127,76 ms. Arsitektur ini memberikan keseimbangan optimal antara performa tinggi dan efisiensi komputasi, sehingga ideal diterapkan pada perangkat pintar seperti ponsel dan sistem IoT dalam konteks manajemen sampah perkotaan. Penelitian ini menegaskan efektivitas transfer learning dalam membangun sistem klasifikasi sampah yang cerdas dan efisien untuk mendukung program pemilahan sampah di tingkat rumah tangga dan institusi.
Usability Test of Mental Health Application MoodPath with Software Usability Measurement Inventory
Full text linkMoodPath is a mobile application for mental health. The application uses Patient Health Questionnaire 9 (PHQ-9) and General Anxiety Disorder 7 (GAD-7) to assess mental health of its users. The study held usability test using Software Usability Measurement Inventory (SUMI) questionnaire with 27 respondents. MoodPath application got usability value of 46.26 that is Below Average in Global SUMI scales. The value is also related with every individual scale in SUMI. The Efficient, Affect, Helpfulness and Control scales have Below Average value. Only the Learnability scale has Above Average value. The usability result is reached with 95% Confidence Interval. Based on the IT skill, respondents with better IT skill gave lower usability score compared to respondent with lesser IT skill. The research also found that familiar UI and standard questionnaire (PHQ-9 and GAD-7) gave positive usability in Learnability and Efficiency scale. The research found that MoodPath application need to consider wider range of users by giving feature that not only satisfied people with lesser IT skill but also people with better IT skill. Based on the usability test, the MoodPath application may improve the usability by providing ‘Remember Me’ and result saving features
Statistical Feature Extraction Based on Wavelet Transform for Arrhythmia Detection
Full text linkEarly detection of arrhythmia through electrocardiogram (ECG) signals is crucial for preventing severe cardiac conditions. This study proposes a binary classification approach using statistical features derived from wavelet-transformed ECG signals. The MIT-BIH Arrhythmia Database was used, with signals filtered using a 0.5–50 Hz Butterworth bandpass filter. Signals were segmented into 360-sample windows with 100-sample overlap and labeled based on the majority annotation within each window. Wavelet transformation using Symlet 8 at level 4 was applied, followed by the extraction of eight statistical features: mean, standard deviation, variance, skewness, kurtosis, interquartile range (IQR), root mean square (RMS), and zero crossing rate (ZCR). These features were classified using MLP, KNN, and SVM models. MLP and KNN achieved the highest accuracy of 92.46%, while SVM had lower accuracy (72.99%) but high recall (94.21%). The results demonstrate the effectiveness of wavelet-based statistical features for lightweight and accurate arrhythmia detection.Early detection of arrhythmia through electrocardiogram (ECG) signals is crucial for preventing severe cardiac conditions. This study proposes a binary classification approach using statistical features derived from wavelet-transformed ECG signals. The MIT-BIH Arrhythmia Database was used, with signals filtered using a 0.5–50 Hz Butterworth bandpass filter. Signals were segmented into 360-sample windows with 100-sample overlap and labeled based on the majority annotation within each window. Wavelet transformation using Symlet 8 at level 4 was applied, followed by the extraction of eight statistical features: mean, standard deviation, variance, skewness, kurtosis, interquartile range (IQR), root mean square (RMS), and zero crossing rate (ZCR). These features were classified using MLP, KNN, and SVM models. MLP and KNN achieved the highest accuracy of 92.46%, while SVM had lower accuracy (72.99%) but high recall (94.21%). The results demonstrate the effectiveness of wavelet-based statistical features for lightweight and accurate arrhythmia detection
Prioritas Pemakaian Anggaran Pada Klinik Kesehatan Berbasis Metode Simple Additive Weighting
Full text linkFinancial management that is less careful and less able to choose each expense will have an impact on the imbalance of income and clinic expenses, if it happens continuously will cause the clinic to gradually be unable to operate anymore. Each type of clinic expense certainly has a different level of importance, the level of importance is influenced by various variables such as the nominal expenditure, deadline date, user of the expenditure, level of importance, purpose of use. The problem that occurs is that the fulfillment of the needs of the clinic is often unable to prioritize expenses or without considering the level of importance of expenses. There needs to be an appropriate priority scale to regulate the level of expenditure that will impact the financial security of the clinic so that it can support the smooth operation of the clinic, with the aim that the clinic will be able to grow or develop for the smooth provision of health services to the community. One method that can be implemented to help prioritize the use of the clinic budget is simple additive weighing. From the results of the comparison between the ranking of the old model and the ranking using the simple additive weighing method, there are two different sequences, namely codes A7 and A14, resulting in a 90% accuracy rate.Improper financial management and the inability to properly manage each expense will lead to an imbalance in the clinic\u27s income and expenses. If this continues, the clinic will eventually become unable to operate. Each type of clinic expense naturally has a different level of importance, influenced by various variables such as the nominal amount of the expense, the time limit for the expense, the user of the expense, the level of importance, and the intended use. The problem that occurs is that meeting the clinic\u27s needs often fails to prioritize expenses and does not consider the level of importance of the expenditure. An appropriate priority scale is needed to regulate the level of expenditure that will impact the clinic\u27s financial resilience so that it can facilitate the smooth operation of the clinic, with the aim of the clinic being able to grow or develop for the smooth provision of health services to the community. One method that can be applied to help prioritize the use of the clinic\u27s budget is simple additive weighting. From the results of trials on several datasets, an accuracy rate of 90% was produced.
Comparative Study of Information System Governance Frameworks: Foundations for IT Risk Management Using COBIT 2019 and ITIL
Full text linkIn this study, COBIT 2019 and ITIL V4 are compared in the context of managing IT risk. Through systematic literature review (SLR), the theoretical and practical foundations of both frameworks are evaluated. COBIT 2019 offers a structured approach, while ITIL emphasizes adaptive operational practices. Analysis of strengths and weaknesses helps organizations choose an approach that aligns with their strategic objectives. With this understanding, organizations can enhance their ability to manage IT risks and achieve business goals effectively
ANALISIS KUALITAS WEBSITE E-GOVERNMENT DENGAN MENGGUNAKAN METODE E-GOVQUAL MODIFIKASI
Full text linkPelayanan Electronic Government yang diberikan oleh pemerintah selalu diharapkan dapat berfungsi dengan baik untuk memberikan kualitas pelayanan yang optimal sesuai dengan tujuan organisasi untuk memenuhi kepuasan masyarakat. Pemerintah Kabupaten Kubu Raya membuat website LAPOR (Layanan Aspirasi dan Pengaduan Online Rakyat), namun laporan masyarakat di website belum sepenuhnya ditindaklanjuti sesuai batas waktu yang ditentukan dan halaman tampilan rating tidak memberikan informasi terkait rating layanan website. Penelitian ini bertujuan untuk menganalisis kualitas website LAPOR Kabupaten Kubu Raya menggunakan metode E-Govqual yang dimodifikasi dengan variabel efisiensi (EF), kepercayaan (TRS), keandalan (RLB), dukungan warga (CS), dan keseluruhan (OVR) pada kepuasan pengguna. Analisis terhadap 100 data responden menggunakan website LAPOR di Kabupaten Kubu Raya menggunakan analisis regresi linier berganda dengan 4 variabel independen pada 1 variabel dependen melalui penyebaran kuesioner. Terdapat 22 hipotesis dalam penelitian ini dengan hasil pengujian yang menyatakan bahwa 7 hipotesis berpengaruh dan berpengaruh signifikan terhadap kepuasan pengguna dengan nilai kualitas 82,6%. Hasil penelitian yang dilakukan adalah memberikan rekomendasi kepada pengelola website LAPOR Kabupaten Kubu Raya untuk memperbaiki sistem pada setiap indikator variabel yang tidak berpengaruh signifikan terhadap kepuasan pengguna. Terdapat 22 hipotesis dalam penelitian ini dengan hasil pengujian yang menyatakan bahwa 7 hipotesis berpengaruh dan berpengaruh signifikan terhadap kepuasan pengguna dengan nilai kualitas 82,6%. Hasil penelitian yang dilakukan adalah memberikan rekomendasi kepada pengelola website LAPOR Kabupaten Kubu Raya untuk memperbaiki sistem pada setiap indikator variabel yang tidak berpengaruh signifikan terhadap kepuasan pengguna. Terdapat 22 hipotesis dalam penelitian ini dengan hasil pengujian yang menyatakan bahwa 7 hipotesis berpengaruh dan berpengaruh signifikan terhadap kepuasan pengguna dengan nilai kualitas 82,6%. Hasil penelitian yang dilakukan adalah memberikan rekomendasi kepada pengelola website LAPOR Kabupaten Kubu Raya untuk memperbaiki sistem pada setiap indikator variabel yang tidak berpengaruh signifikan terhadap kepuasan pengguna.The Electronic Government services provided by the government are always expected to function properly to provide optimal service quality according to organizational goals to meet community satisfaction. The Kubu Raya Regency Government created LAPOR website (People\u27s Online Aspirations and Complaints Service), but community reports on website have not been fully followed up according to specified deadline and the rating display page does not provide information regarding website service ratings. This study aims to analyze quality of the Kubu Raya Regency LAPOR website using the modified E-Govqual method with variables of efficiency (EF), trust (TRS), reliability (RLB), citizen support (CS), and overall (OVR) on user satisfaction. Analysis of 100 data from respondents using the LAPOR website in Kubu Raya Regency used multiple linear regression analysis with 4 independent variables on 1 dependent variable through distributing questionnaires. There are 22 hypotheses in this study with the test results stating that 7 hypotheses have an influence and significant effect on user satisfaction with a quality value 82.6%. The results of the research conducted were to provide recommendations to the Kubu Raya Regency LAPOR website manager to improve the system on each indicator of the variables that have no significant effect on user satisfaction
Optimasi Clustering K-Means Menggunakan Algoritma Genetika Dengan Data View Dan Like Di Tiktok
Full text linkK-Means merupakan algoritma yang sering digunakan untuk melakukan pengelompokkan atau sering juga disebut clustering. Dengan menentukan pusat centroid awal secara random pada algoritma K-Means akan ditingkatkan performanya menggunakan Algoritma Genetika (GA).
Menggunakan data set publik di Kaglle, berupa data set tiktok dimana jumlah view dan like dengan record data sebanyak 19.084 setelah dilakukan pembersih data. Yang akan diuji dengan melakukan performa clustering K-Means dengan Algoritma Genetika. Dan untuk validitas nya nanti menggunakan Davis Boulden Index, dimana hasil validitas DBI ini nanti akan meningkatkan performance K-Means dengan menambahkan
Algoritma Genetika.
Dengan pengujian K-Means dengan jumlah k=3, k=4 dan k=5 menghasilkan masing-masing validitas DBI 0,64 ; 0,79 dan 0,72. Sedangkan untuk algoritma K-Means dengan peningkatan performa
menggunakan GA didapatkan validitas dengan masing-masing DBI sebagai berikut 0,45 ; 0,40 dan 0,60. Dengan hasil penelitian menghasilkan bahwa peningkatan performa K-Means dengan menggunakan GA memberikan hasil validitas lebih kecil dari pada hanya menggunakan perhitungan KMeans saja
Komparasi Metode SVM dan Adaboost untuk Klasifikasi Kanker Payudara
Full text linkOne of the most prevalent malignancies in women and a major global cause of death is breast cancer. To determine whether a cancer is benign or malignant, early detection is essential. The usefulness of the Support Vector Machine (SVM) and Adaptive Boosting (Adaboost) algorithms for breast cancer classification using mammography data is compared in this study. 569 records make up the dataset, which was sourced from the Kaggle Repository and is split into 75% training data and 25% testing data. Preprocessing steps include feature and target variable creation, categorical-to-numerical conversion, data splitting, and normalization. SVM achieved an accuracy of 97%, with a precision of 98%, recall of 94%, and F1 score of 96%. Adaboost, on the other hand, achieved an accuracy of 96%, precision of 98%, recall of 92%, and F1 score of 95%. The results reveal that both algorithms are highly effective for breast cancer detection, with SVM marginally exceeding Adaboost in total performance. These findings emphasize the promise of machine learning techniques in facilitating early cancer diagnosis, hence boosting survival rates. It is advised that future research employ a wider range of datasets and investigate different classification techniques in order to improve accuracy and dependability even more.One of the most prevalent malignancies in women and a major global cause of death is breast cancer. To determine whether a cancer is benign or malignant, early detection is essential. The usefulness of the Support Vector Machine (SVM) and Adaptive Boosting (Adaboost) algorithms for breast cancer classification using mammography data is compared in this study. 569 records make up the dataset, which was sourced from the Kaggle Repository and is split into 75% training data and 25% testing data. Preprocessing steps include feature and target variable creation, categorical-to-numerical conversion, data splitting, and normalization. SVM achieved an accuracy of 97%, with a precision of 98%, recall of 94%, and F1 score of 96%. Adaboost, on the other hand, achieved an accuracy of 96%, precision of 98%, recall of 92%, and F1 score of 95%. The results reveal that both algorithms are highly effective for breast cancer detection, with SVM marginally exceeding Adaboost in total performance. These findings emphasize the promise of machine learning techniques in facilitating early cancer diagnosis, hence boosting survival rates. It is advised that future research employ a wider range of datasets and investigate different classification techniques in order to improve accuracy and dependability even more.