Journal of Statistics and Data Science

Comparison of Poverty Clustering Results based on Distance Measurement with the Complete Linkage Method in Indonesia

2025-04-17T07:33:39+00:00

Every year, population growth in Indonesia increases and has the potential to trigger poverty.
Poverty indicators include the number of poor people, per capita expenditure, human
development index, average years of schooling, and unemployment. The clustering of regions
is necessary for the government to be more effective in development. One of the methods
used is cluster analysis, a statistical technique that groups objects based on similar
characteristics. This research compares the results of clustering poverty in Indonesia's
Regency/City in 2023 using the complete linkage method, which is based on the farthest
distance. The distances analyzed include Euclidean, Square Euclidean, Manhattan, and
Minkowski, resulting in two clusters at each distance. Minkowski proved to be the best
distance with the smallest standard deviation ratio, which was 1.518 for cluster 1 and 2.225
for cluster 2, compared to the other distances. These results show that the Minkowski method
is superior in clustering poverty areas in Indonesia.

Machine Learning Approach to Automated Early Warning System for Medical Vital Signs Monitoring

2025-03-05T14:30:41+00:00

Precise and timely detection of deteriorating vital signs is an important aspect of patient safety and clinical intervention. The current standard of monitoring systems lacks automated early warning systems, instead using manual observation to make judgments. This manual approach can lead to delays in detecting critical changes in a patient's condition. We present a novel approach to developing an automated early warning system for vital signs using a hybrid method that combines LSTM (Long Short Term Memory) and XGBoost (Extra Gradient Boost), both methods offer robust predictive modeling that is able to capture the complex and often non-linear relationships inherent in physiological data. This research believes that using a novel technique that combines LSTM and XGBoost advances predictive systems in healthcare-based technology as well as laying the groundwork for even further innovations in early warning systems. The early warning system will evaluate vital signs such as respiratory rate, SpO₂ levels, heart rate, body temperature, and pulse which can recognize and predict early signs of clinical deterioration, allowing early intervention and may save a patient’s life. This research will use error metrics such as MAPE, MAE (Mean Absolute Error), MSE, RMSE, and MAD to compare the predicted actual values.

The ROCK Ensemble Cluster Method for People's Welfare Analysis A Mixed Data Approach

2025-03-25T03:00:49+00:00

This study clusters districts/cities in West Sumatra based on public welfare indicators using the Ensemble Cluster Method with the ROCK algorithm. This approach handles mixed data, where numeric data is clustered with Hierarchical Agglomerative Clustering, while categorical data uses ROCK. The clustering results are combined through Cluster Ensemble to improve accuracy. Secondary data from BPS 2023 includes eight indicators of people's welfare. Clustering was validated using Compactness (CP). Results showed five optimal clusters, with a CP value of 0.44. Cluster 1 has the greatest welfare challenges, while Cluster 5 shows the highest welfare. These findings can be used as a basis for formulating more targeted regional development policies.

Application of Tobit Regression on Household Expenditure on Egg and Milk Consumption in Bengkulu City

2025-03-07T02:05:06+00:00

Regression analysis is a statistical method used to examine the functional relationship between two or more independent variables and a dependent variable. One of the regression methods designed to handle censored data or data with significant zero values is Tobit regression. This study aims to model household expenditures on egg and milk consumption in Bengkulu City using Tobit regression and to identify the factors influencing these expenditures. The data were obtained from the 2022 National Socioeconomic Survey, with a total sample of 1,170 households. The Tobit regression model was chosen because most household expenditure data had zero values, indicating censored data characteristics. This study identified several factors affecting expenditures on egg and milk consumption, such as the household head's education level, the number of household members, and the household head's employment sector. The results showed that the education level of the household head (elementary, junior high, and high school), the number of household members, and the household head's employment in agriculture and trade sectors had significant impacts on household expenditures for egg and milk consumption. The education level of the household head and their employment sector had a negative relationship, while the number of household members showed a positive relationship with these expenditures. The Tobit regression model successfully modeled household expenditures with adequate accuracy, as indicated by a Mean Absolute Percentage Error (MAAPE) of 1.38%.

Enhancing Bus Scheduling Efficiency in Rajshahi City, Bangladesh Through Linear Programming Apporach

2025-03-24T03:06:58+00:00

Efficient bus scheduling is significant for optimizing transportation systems, particularly in densely populated areas like Rajshahi division in Bangladesh. This research focuses on minimizing the no of buses needed for operations in Rajshahi Bus Terminal by introducing a rescheduling strategy using Linear Programming Approach. The study divides the day into eight time periods, with four overlapping shifts per day. By strategically overlapping shifts, we aim to utilize existing resources more efficiently and reduce the number of buses needed. Utilizing mathematical modeling formulation of Linear Programming and optimization techniques, an optimal rescheduling method is introduced to minimize the number of buses required ensuring efficient service delivery and passenger satisfaction. Furthermore, it evaluates the effectiveness of the proposed approach through quantitative analysis and compares it with the remained scheduling system in terms of bus utilization and overall efficiency. The outcomes of this research contribute to the optimization of bus transportation systems in urban areas, offering insights into effective resource allocation. By implementing the suggested re-scheduling strategy at Rajshahi city Bus Terminal, public transportation networks can be made more sustainable and cost effective. This could serve as a model for other transportation hubs in Bangladesh and beyond.