Publication Search

Accurate rainfall prediction is needed to improve the performance of land that always uses rainfall data. Data mining or often called knowledge discovery in databases (KDD) is an activity that includes collecting, using historical data to find regularities, patterns or relationships in large data. In predicting rainfall, there are several conditions that can be observed as reference data to predict rainfall, namely wind speed, temperature, and air humidity. In this research, a backpropagation artificial neural network prediction method is developed that can be used in predicting future rainfall. The backpropogation artificial neural network method that was built produced an accuracy value of 95.36%, a precision value of 90.50%, a recall value of 97.50% and an f-measure value of 92.00%

The main objectives are to identify emotion patterns hidden in K-Pop music based on audio features extracted from the Spotify API and to build an emotion classification model that can predict the emotions of K-Pop songs.In this approach, the K-Means algorithm is used to cluster K-Pop songs based on audio features such as energy, valence, tempo, danceability, and speechiness. The clustering results reveal several main groups that represent variations in musical characteristics and emotions. Next, the C4.5 algorithm was used to build an emotion classification model based on the clustering results. The C4.5 model showed high performance with accuracy reaching 99.48% on a 90:10 dataset split, 99.21% on an 80:20 split, and 98.95% on a 70:30 split.The Streamlit application was developed to visualize emotion predictions from K-Pop songs with a web-based user interface. In addition, Ngrok was used to provide remote access to this application, allowing users to test and use the application remotely.The results of this study show that the combination of K-Means and C4.5 can effectively cluster and classify emotions in K-Pop music, providing valuable insights into the musical characteristics that influence emotions. This application has the potential to be used in further analysis, development of intelligent features in music applications, and improvement of user experience in listening to K-Pop music.

This study aims to apply the K-Nearest Neighbors (KNN) algorithm to predict stunting status in young children based on height and weight data. Stunting is a growth failure condition caused by chronic malnutrition that negatively impacts children's physical and mental development. The dataset includes height, weight, and stunting status of children. The results show that the KNN model with k=3 achieved 100% accuracy on the test data. Evaluation using the confusion matrix and classification report indicates perfect precision, recall, and F1-score for each class. Data normalization with StandardScaler improved the model's performance by ensuring all features are on the same scale. The KNN algorithm proves to be a simple yet effective method for predicting stunting, demonstrating significant potential for early detection and health intervention in children. This study recommends using a larger and more diverse dataset, as well as incorporating additional relevant features to enhance model accuracy. Implementing the model in a web or mobile application is also suggested to assist healthcare professionals in the field.

This research aims to determine whether there are differences in bankruptcy prediction result between the Altman model and the Springate model for building construction sub-sector companies listed on the Indonesia Stock Exchange. The type of research used comparative quantitative research. The population used is building construction sub-sector companies listed on the Indonesia Stock Exchange during the years 2018-2022. Sample collection uses purposive sampling technique. The analysis technique uses the Wilcoxon signed test. The Wilcoxon signed test indicate that the Altman and Springate models show significant difference in bankruptcy prediction result for building construction sub-sector companies listed on the Indonesia Stock Exchange.

Forecasting demand for screen-printed clothing products at the UMKM "D'mitz Screen Printing" in Sobrah Village, Wungu District, Madiun Regency helps with production control planning to maximize supply chain management for screen-printed clothing products. To predict future product demand, it is very important for UMKM to forecast market demand. Forecasting future demand is very important to avoid sales prediction errors that can cause waste, such as increased production costs due to sales predictions being too large, or stock outs due to sales predictions being too small, which results in customers having to wait longer to get the goods they want. Based on this problem, the UMKM "D'mitz Screen Printing" carried out a demand forecasting analysis for screen printed clothing with the aim of reducing waste and maximizing value. Forecasting demand for screen printed clothing for the next five months using time series analysis and moving average methods. Forecasting results for the period March 2022 to February 2023 show sequential forecasting values of 3266.67; 3300; 3250; 3283.33; 3233.33; 3316.67; 3333.33; 3372.22; 3305.56; and 3272.22. From the Mean Absolute Error (MAE) and Mean Square Error (MSE) calculations that have been carried out, the MAE value is 94.44 and the MSE value is 16018.593.

This research explores various machine learning approaches, including deep learning and ensemble methods, to predict climate change indicators. We focus on temperature and precipitation trends using large datasets spanning multiple decades. By comparing the performance of algorithms like CNN, RNN, and random forests, we identify the most accurate models for specific climate variables. Our findings demonstrate that ensemble models provide better accuracy and reliability, especially for temperature predictions.

Abstract: Renewable Energi have been goverment concern in Indonesia since 2014. HHO generator is one of some machine to generate hydrogen fuel using electrolysis process. In this research, the discussion will focus on the comparison between theoretical predictions and experimental results, about the causes of precipitation and corrosion of electrodes made of stainless steel produced from HHO generators, especially when using a mixture of KOH and NaCl electrolytes. The results, show the precipitate produced by the KOH solution is less compared to the NaCl solution. Meanwhile, the chlorine (Cl⁻) content of NaCl which is initially corrosive is not always present because it is formed into Cl₂ in the form of a gas, so the subsequent precipitation and corrosiveness depend on the element hydroxide (OH⁻) from H₂O dissociation.

Big data has revolutionized the landscape of natural sciences by providing extensive datasets that enable deeper insights and more accurate predictions. However, effectively analyzing such vast and complex data requires optimized machine learning algorithms tailored to specific applications. This study focuses on enhancing the performance of machine learning models in big data analysis for applications in natural sciences. The research aims to identify key optimization techniques, including feature selection, hyperparameter tuning, and algorithm customization, to improve model accuracy and computational efficiency. A combination of supervised and unsupervised learning approaches was applied to real-world datasets in fields such as climate science, genomics, and ecology. The findings demonstrate significant improvements in predictive accuracy and processing speed, highlighting the potential of optimized machine learning techniques in solving complex problems in natural sciences. The implications of this research extend to more efficient resource utilization and improved decision-making in scientific exploration and environmental management.

Artificial intelligence (AI) is transforming paediatric diabetes management, offering innovative solutions for monitoring, treatment, and prediction. This mini-review explores how AI is being utilised to improve the care of children with diabetes mellitus, focusing on its application in glucose monitoring systems, predictive algorithms, and personalised treatment plans. The study synthesises recent advancements in AI technologies, examining their impact on enhancing the accuracy of diagnosis, reducing the burden on healthcare providers, and improving patient outcomes. Through a systematic review of the literature, key AI tools and models that have shown promise in paediatric diabetes care are identified. The findings highlight the potential of AI to revolutionise diabetes management, with implications for both clinical practice and future research. However, challenges remain in ensuring the ethical implementation and integration of these technologies into existing healthcare systems. The paper concludes with recommendations for advancing AI applications in this field, emphasising the need for continued innovation and collaboration between healthcare professionals and AI developers.

Predicting sales is an important aspect in sales development. Sales prediction simulation is an estimated calculation of the level of product sales in a certain period. Sales of pempers tend to fluctuate due to the choice of many brands available, resulting in sales having a little difficulty in estimating sales of MAKUKU diaper products at Greens Mart stores. This research aims to predict sales Pempers products. This research uses several stages of identifying problems, determining research objectives, collecting data, data collected or obtained from interviews with salespeople, managing data using Monte Carlo stages, implementing/testing data and testing results to see the accuracy of the method used. The analysis results show that Comfit M and Comfit L have almost the same level of accuracy, namely Comfit M is around 90.63% and sales of Comfit L are 90.48%. These values ​​provide an indication of the level of accuracy of the sales predictions made.      

This study employed knowledge discovery in databases (KDD) to extract and discover knowledge from the Benue State Polytechnic (Benpoly) admission database and used a structural causal model (SCM) ontological framework to represent the admission process in the Nigerian polytechnic education system. The SCM ontology identified important causal relations in features needed to model the admission process and was validated using the conditional independence test (CIT) criteria. The SCM ontology was further employed to identify and constrain input features causing bias in the local interpretable model-agnostic explanations (LIME) framework applied to machine learning (ML) black-box predictions. The ablation process produced more stable LIME explanations devoid of fairness bias compared to LIME without ablation, with higher prediction accuracy (91% vs. 89%) and F1 scores (95% vs. 94%). The study also compared the performance of different ML models, including Gaussian Naïve Bayes, Decision Trees, and Logistic Regression, before and after ablation. The limitation is that the SCM ontology is qualitative and context-specific, so the fair-LIME framework can only be extrapolated to similar contexts. Future work could compare other explanation frameworks like Shapley on the same dataset. Overall, this study demonstrates a novel approach to enforcing fairness in ML explanations by integrating qualitative SCM ontologies with quantitative ML/LIME methods.

Rotary drilling is one of the drilling methods that is often used for stripping overburden in open pit mining activities. In the rotary drilling process, prediction of the penetration rate is very important for scheduling production and estimating drilling costs. Drilling in mining activities is used, among other things, to create blast holes. In blasting activities, drilling is the first activity that must be carried out to provide a blast hole which will later be filled with explosives to be detonated. Meanwhile, the speed at which the blast hole is prepared is influenced by the speed of the drilling tool to penetrate the rock. Drilling speed is influenced by two factors, namely internal factors and external factors.      

This research discusses the implementation and evaluation of the Convolutional Neural Network (CNN) convolutional neural network model for classification of fruit types, specifically to differentiate between Banana and Papaya. The CNN model used consists of several convolutional, pooling, and fully connected (dense) layers designed to extract features and perform binary classification. Data augmentation is applied to the training set to increase data variation and prevent overfitting. The image data used is normalized to speed up training convergence. The model was trained using the Adam optimizer and the binary crossentropy loss function for 20 epochs. Performance evaluation was carried out using the validation set. The results show that the model is able to effectively classify fruit images with a high level of accuracy. Predictions are made by uploading images, resizing them, and normalizing them before using the model for predictions. The classification threshold was set at 0.4, where a predicted probability greater than or equal to 0.4 was classified as Banana and a probability less than 0.4 was classified as Papaya. This research shows that the CNN model can be used effectively for binary image classification tasks and can be extended to classify more types of fruit with appropriate data adjustments and model architecture.

Community Health Centers (Puskesmas) are the main place for the public to get basic health services. In certain conditions, the Health Center will receive visits from patients who arrive with the same number of origins at the same time. This has an impact on health services experiencing a decline and not maximizing the health of health workers in checking patients' health. So, comprehensive services are not obtained, and there are even patients who queue for too long. In order to minimize this occurrence, simulations are carried out to predict the number of possible visits that will occur in the future. The method used in this study is Monte Carlo. This research is intended so that the information provided by the Health Center related to the prediction of the number of patient visits that are likely to occur in the next year or in the future can be done validly and accurately. The data used is data on the number of patient visits in 2021, 2022, and 2023. As a result of this study, a prediction of the number of patient visits was obtained so that the Health Center could implement the next action that was useful for improving the quality of services needed    

Stunting is a condition of failure to thrive in children, in Indonesia it is still a serious problem with a fairly high prevalence. The government is trying to reduce stunting rates with various health programs, and early detection through routine measurements is very important. This research uses the Extreme Gradient Boosting (XGBoost) algorithm to classify stunting status in children under five years. This study uses a relevant dataset containing anthropometric information on children, such as gender, age, birth weight and length, current weight and length, and breastfeeding status. The research stages include dataset search, preprocessing, classification, evaluation, and implementation in a local web-based prediction program. The XGBoost algorithm was chosen because of its advantages in speed, scalability, and efficiency. After preprocessing and data sharing, the model was trained and tested, resulting in 86% accuracy, 89% precision, 95% recall, and 92% F1-score. Evaluation using the confusion matrix and classification report shows that this model is quite effective in classifying stunting status.  

Research on forecasting new student admissions at SD Negeri 10 Koto Tinggi Surian employs data analysis using simple linear regression to predict the number of new students based on previous years and formulate it into a mathematical model. In its calculation process, this approach can adopt both quantitative and qualitative methods. The prediction error rate for the next year is very low. Out of 41 predicted students, only one differs from the actual number. The study's data spans student admissions from 2019 to 2023. The research findings demonstrate the effectiveness of conventional linear regression techniques. The prediction for 2023 indicates an intake of 42 students. Therefore, it can be concluded that simple linear regression accurately predicts new student admissions at SD Negeri 10 Koto Tinggi Surian with high accuracy.    

This research discusses the prediction of coffee production outcomes in Kalimantan using Markov chain analysis. A Markov chain is an analytical technique that can be used to predict future changes based on past changes. The aim of this study is to determine the predicted coffee production results in Kalimantan from 2023 to 2025 based on Markov chain analysis. Based on the results of the Markov chain analysis of coffee production data in Kalimantan from 2018 to 2022, it is concluded that the predicted coffee production results in 2024 are as follows: West Kalimantan Province is expected to produce 3.5632 thousand tons, Central Kalimantan Province 0.3275 thousand tons, South Kalimantan Province 1.3727 thousand tons, East Kalimantan Province 0.1934 thousand tons, and North Kalimantan Province 0.1539 thousand tons. Furthermore, in 2025, the coffee production in West Kalimantan Province is predicted to reach 3.5935 thousand tons, Central Kalimantan Province 0.3304 thousand tons, South Kalimantan Province 1.3857 thousand tons, East Kalimantan Province 0.1952 thousand tons, and North Kalimantan Province 0.1554 thousand tons.

Basnah, Nim.19317001005. Penerapan pembelajaran prediction guide terhadap kemampuan menulis teks resensi pada siswa kelas XI SMA Negeri 1 Tanoh Alas Tahun Pembelajaran 2023/2024. Penelitian ini bertujuan untuk mengetahui Penerapan pembelajaran prediction guide terhadap kemampuan menulis teks resensi pada siswa kelas XI SMA Negeri 1 Tanoh Alas Tahun Pembelajaran 2023/2024. Adapun yang menjadi populasi dalam penelitian ini adalah seluruh siswa kelas XI yang berjumlah 22 orang dan sampel berjumlah 22 orang. Metode yang digunakan dalam penelitian ini adalah metode eksperimen dengan rancangan one group pretest and postest design dengan analisis data menggunakan uji “t”. Dari pengelolahan data di peroleh nilai rata-rata tes awal kemampuan menulis teks resensi oleh siswa sebesar 67,73 dan termasuk dalam kategori kurang, dengan standar deviasinya sebesar 2,91 dan mengalami peningkatan nilai hasil tes akhirnya dengan nilai rata-rata sebesar 81,14 dan termasuk dalam kategori baik, dengan standard deviasinya sebesar 2,10. Berdasarkan hasil nilai tes awal dan tes akhir siswa tersebut dan jika di kaitkan dengan nilai KKM mata pelajaran bahasa Indonesia untuk SMA sebesar 75 maka kemampuan menulis teks resensi siswa termasuk dalam kategori baik. Dari pengujian hipotesis diperoleh nilai t hitung = 17,64 dan di konsultasikan dengan nilai  pada taraf signifikan 5% = 2,080 dengan demikian > atau 17,64> 2,080 maka hipotesis di terima. Jadi dapat disimpulkan bahwa ada Penerapan pembelajaran prediction guide terhadap kemampuan menulis teks resensi pada siswa kelas XI SMA Negeri 1 Tanoh Alas Tahun Pembelajaran 2023/2024.

This research utilizes the Markov Chain method to analyze daily weather data in the city of Medan. The main objective of this study is to forecast weather changes in the future based on the weather conditions of the previous day. Daily weather data was collected from the nearest weather station over a specific period of time. The analysis results indicate that the Markov Chain model provides good estimates of the likelihood of weather changes from one day to the next. The steady state probabilities demonstrate the dominance of partly cloudy and clear weather in the long term. This research provides valuable insights for various sectors related to weather, such as agriculture, transportation, and tourism.

Indonesia, as a country with complex geological conditions due to the convergence of various tectonic plates, is highly susceptible to natural disasters such as earthquakes, tsunamis, and volcanic eruptions. The city of Semarang, as the capital of Central Java Province, also frequently faces disasters such as floods, landslides, and earthquakes. Predicting the occurrence of natural disasters becomes crucial to mitigate the negative impacts they cause. This study uses the Markov chain method to predict natural disasters in the city of Semarang based on disaster data from 2018-2022. The prediction results indicate a 16% chance of floods, 34% chance of landslides, 10% chance of tornadoes, 22% chance of fires, and 17% chance of falling trees in 2023. Validation of the predictions against actual data for 2023 shows a relatively good match for floods and fires, but there are significant differences in the predictions for tornadoes and falling trees. These results indicate that the Markov chain method has potential in predicting disaster occurrences, but accuracy improvements are needed to account for weather variability and dynamic environmental factors. This research is expected to assist the government and society in enhancing disaster preparedness and mitigation in the future.