Triastuti Wuryandari
Departemen Statistika, Fakultas Sains Dan Matematika, Universitas Diponegoro

Published : 32 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 32 Documents
Search

PERBANDINGAN KINERJA METODE KLASIFIKASI K-NEAREST NEIGHBOR DAN SUPPORT VECTOR MACHINES PADA DATASET PARKINSON Ridho, Wahyu Anwar; Wuryandari, Triastuti; Hakim, Arief Rachman
Jurnal Gaussian Vol 12, No 3 (2023): Jurnal Gaussian
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/j.gauss.12.3.372-381

Abstract

The government program in the form of social assistance (bansos) is part of the effort to improve the welfare of the community and ensure basic needs and improve the standard of living of the recipients. However, there are often cases of mistargeting of social assistance programs by the government. Improper data management and Data Terpadu Kesejahteraan Sosial (DTKS) which are not used as the cause of the distribution of social assistance are not well targeted. The data can be analyzed using the classification method to determine whether or not the family accepts the ban from the government. This study classifies the SUSENAS data by comparing K-Nearest Neighbor (KNN) and Support Vector Machines (SVM). The advantage of the KNN method lies in the level of accuracy to solve problems with large data while the SVM method has better performance in various fields of application such as bioinformacs, handwriting recognition, text classification and so on. Based on training data and testing data comparison 85%:15% showed that KNN method had a better classification performance than the SVM method. The accuracy value of KNN method is 80,95% higher than the accuracy value of SVM method is 78,79%.
PERAMALAN INDEKS HARGA SAHAM GABUNGAN (IHSG) MENGGUNAKAN MODEL INTERVENSI FUNGSI PULSE Rosilawati, Elsa Dwi; Tarno, Tarno; Wuryandari, Triastuti
Jurnal Gaussian Vol 12, No 3 (2023): Jurnal Gaussian
Publisher : Department of Statistics, Faculty of Science and Mathematics, Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.14710/j.gauss.12.3.382-391

Abstract

The intervention model is one model that is frequently used to explain how interventions from both internal and external sources can lead to dramatic fluctuations in a time series of data. The Composite Stock Price Index, known as the IDX Composite, is an index that tracks all stock price performance. For the Composite Stock Price Index from 2 October 2020 to 6 June 2022, daily close price data are used in this study. The data showed a sharp reduction starting on 9 May 2020 (T=386) and lasting for the following 4 days, which made the pulse function the likely intervention model. Rising interest rates and high inflation figures from the United States are to blame for the drop in IDX Composite close price. In addition, a lot of profit-taking was done because of the Eid holidays and the expectation of a substantial increase in COVID-19. The best intervention model created is ARIMA ([3],1,0) with an intervention order of b=0, r=0, and s=11, which can then be used to forecast Composite Stock Price Index for the following period. This is based on the outcomes and analyses. The sMAPE value in the research utilizing this model was 0.98%, suggesting very strong forecasting capabilities.