Enthusiastic : International Journal of Applied Statistics and Data Science

K-Means Clustering Application of Open ‎Unemployment in 2020 Caused by COVID-19 in West Java Province

M. Ficky Haris Ardiansyah — 2024-04-28

West Java was the province with the highest unemployed rate during the COVID-19 pandemic. Significant increase of open ‎unemployment rate in West Java negatively impacts the national income. This study aims to apply the ‎clustering method using the k-means algorithm to determine priority clusters in West Java ‎Province by looking at the number of clusters in West Java’s city and the main characteristic of ‎each cluster. The clustering was conducted utilizing a k-means clustering algorithm which is grouping data based on similar ‎characteristics. The clustering results were evaluated using silhouette method. The results indicated that ‎two clusters were optimal. The clustering process using the k-means method showed that there were three clusters distinguishing the open unemployment rate during the pandemic in West Java Province in 2020. Cluster 1 ‎had a fairly low open unemployment rate due to the stalled service sector and low minimum city wage. ‎Cluster 2 had a high open unemployment rate due to the service sector and high minimum city wage. ‎Cluster 3 had medium open unemployment rate due to the service sector and also medium minimum city ‎wage. It suggests that cluster 2 is a priority cluster in dealing with the open unemployment rate.‎

Performance of Three-Parameters Dirichlet Universal Portfolio During COVID-19 Pandemic

Goh Yeok Qin — 2024-04-28

Stock returns are often the primary objectives for investors, financial analysts as well as the politicians with the intention to make a right investment decision. In this paper, we study the performance of the three-parameters Dirichlet universal portfolio on selected stocks during the COVID-19 pandemic. Some empirical results are obtained based on some selected data sets from the local stock exchange. The period of trading of the stocks are selected from 2nd January 2020 to 18th August 2021 consisting of 400 trading days. The empirical results seem to indicate the three-parameters Dirichlet universal portfolio performs well during the COVID-19 pandemic by a proper choice of parameters. Also, this study provides evidence that the capital achievement at the end of the 400th trading days is influenced by the arrangement of the stocks within each selected data set. Besides, the performance of the homogeneous datasets, particularly, main data set from healthcare sector, is better than heterogeneous datasets during the COVID-19 pandemic.

Characterization of Student’s Performance in Massive Open Online Courses (MOOC)

Tan Ching Joe — 2024-04-28

Massive Open Online Courses (MOOC) allow students to learn online at any time and from any location. Unfortunately, poor completion rates and a large student group make it difficult for teachers to keep track of their student’s progress. Due to a lack of adequate counselling, students who perform poorly are more likely to give up. The goal of this study was to predict student’s certification by analyzing data on student’s learning behavior. The initial data on learning behavior was obtained from edX, a well-known MOOC platform. Based on this data, three statistical models such as logistic regression, graph convolutional network, and cluster analysis were utilized to predict student’s performance. The proposed model’s usefulness was demonstrated by using a testing set of data from the actual courses. Our findings showed that tracking student activity in terms of number of unique days active, watching videos, participating in forum discussions, and exploring more courseware content might help predict student’s performance in MOOC and enhance completion rates.