Awareness of seismicity activity rates could be learned from modeling the earthquake events by utilizing the record of seismicity events data in NTB over time which is associated with count time series data. Poisson Hidden Markov Model (PHMM) has been widely applied in various fields, including earthquake event. Therefore, it would be interesting to implement PHMM on earthquake case in NTB. The data can be analyze using PHMM as we can ignore the over-dispersion and dependency relationship among data. The model is the development of Markov Model that consists of (a) observed state, which can be observed directly and (b) hidden state, which cannot be observed directly because it is hidden. Hidden state in this research is defined as seismicity activity rates classified into a low rate and high rate (2 states). The count time series data of earthquake events will be more informative when it is classified into the seismicity activity levels. This research applied earthquake event data (magnitude ≤ M4.7 and depth < 60 km) from January 2009 until September 2018, collected from USGS (United States Geological Survey). The parameter estimation method used in this research is the Bayesian method. The objective of this research is to obtain parameters of 2 state Poisson Hidden Markov Model using the Bayesian approach. Model validation measured by MAE (Mean Absolute Error). Based on the result, the average earthquake cases caused by low seismicity activity rate in NTB over time is 1 event whilst the high rate is 18 events. The probability of low seismicity activity rate influenced by the previous rate and the long run behavior (steady state) in NTB is still larger than the high rate. The achieved two-state PHMM is suitable for modeling the earthquake case in NTB with MAE values of 0.4079.
31 May 2025 in Mathematical Modelling of Engineering Problems
This study aims to determine the implementation of Fuzzy Sugeno in classifying textual data obtained from Twitter so as to determine the polarity of public opinion regarding PPKM policies and Covid-19 vaccinations. This study uses primary data via Twitter related to COVID-19 vaccination and PPKM policies in Indonesia starting from February 9, 2021 to January 17, 2022. There are several stages carried out, namely data collection, data pre-processing, data labeling, data weighting. , identification of membership functions, determination of fuzzy sets, formation of classification systems, and evaluation of classification results. The results of this study explain that Fuzzy Sugeno's performance in classifying tweets is quite good with an average accuracy of 89.13%. Meanwhile, public opinion regarding PPKM policies and Covid-19 vaccinations tends to be balanced with 36.92% of tweets classified as positive sentiments, 22.85% negative sentiments, and another 40.23% classified as neutral sentiments. In addition, the fuzzy set that is formed based on the data observation method is very well done because it is able to adjust the frequency of the data in each category. This really helps improve the performance of the built classification system.
1 January 2025 in International Journal of Mathematics and Computer Science
The machine learning technique is a computational algorithm that is currently developing and is widely used in various models because it has the advantage of predicting without limited assumptions. Therefore, the purpose of this research is to develop a machine learning method on the kriging interpolation method to overcome the unmet assumptions of stationary and normally distributed rainfall data in East Java. This research is expected to make an important contribution as input for flood hydrograph models, which will improve understanding of flood behavior and support flood early warning systems in East Java.
7 October 2022 in JTAM (Jurnal Teori dan Aplikasi Matematika)
This study aims to model Principal Component Regression (PCR) using Variational Bayesian Principal Component Analysis (VBPCA) with Ordinary Least Square (OLS) as a method of estimating regression parameters to overcome multicollinearity at various levels of the proportion of missing data. The data used in this study are secondary data and simulation data contaminated with collinearity in the predictor variables with various missing data proportions of 1%, 5%, and 10%. The secondary data used is the Human Depth Index in Java in 2021, complete data without missing values. The results indicate that the multicollinearity in secondary and original data can be optimally overcome as indicated by the smaller standard error value of the regression parameter for the PCR using VBPCA method which is smaller and has a relative efficiency value of less than 1. VBPCA can handle the proportion of missing data to less than 10%. The proportion of missing data causes information from the original variable to decrease, as evidenced by immense MAPE value and the parameter estimation bias that gets bigger. Then the cross validation (Q^2 ) value and the coefficient of determination (adjusted R^2 ) are get smaller as the proportion of missing data increases.
12 June 2019 in Pakistan Journal of Statistics and Operation Research
This research is a development from previous research that has studied the method of spatio temporal disaggregation with State space and adjusting procedures for predicting hourly rainfall based on daily rainfall (Astutik et al, 2013). However, this study is limited to predicting hourly rainfall in some sampled locations in the future. Astutik et al (2017, 2018) have modeled hourly and daily rainfall using posterior predictive bayesian VAR at the Sampean watershed of Bondowoso. This study aims to predict hourly rainfall data based on daily rainfall data in the future at the outsampled locations using posterior predictive bayesian VAR and adjusting procedures in the method of spatio temporal disaggregation.