Empirical Best Linear Unbiased Predictor (EBLUP) has been widely used to predict parameters in area with small or even zero sample size. The problem is when this model should be used to predict the parameters of non-sampled area. Ordinary EBLUP predicted the parameters using synthetic model which ignore the area random effects because lack of non-sampled area information. Thus, those prediction will be distorted based on a single line of the synthetic model. One of idea that developed in this paper is to modify the prediction model by adding cluster information by assuming that there are similiarities among particular areas. These information will be added into the model to modify the intercept of prediction model. Another approach is by adding random effects of auxiliary variable into the previous model in order to modify both intercept and slope of the prediction model. In this paper, simulation process is carried out to study the performance of the proposed models compared with ordinary EBLUP. All models evaluated based on the value of Relative Bias (RB) and Relative Root Mean Squares Error (RRMSE). The results show that the addition of cluster information can improve the ability of the model to predict on non-sampled areas.

The presidential election is one of the political events that occur in Indonesia once in five years. Public satisfaction and dissatisfaction with political issues have led to an increase in the number of political opinion tweets. The purpose of this study is to examine the performance of the k-means and k-medoids method in the Twitter data and to tweet about the presidential election in 2019. The data used in this study are primary data taken from Muhyi's research (2019), then mining the text against data obtained. Because this data has been processed by Muhyi (2019) to analyze the electability of the 2019 presidential candidate pairs, for this journal needs a preprocessing was carried out to analyze the tendency of tweets to side with the candidate pairs of one or two. The difference in the pre-processing of this research with previous research is that there is a cleaning of duplicate data and normalizing. The results of this study indicate that the optimal number of clusters resulting from the k-means method and the k-medoid method are different.

National Education Standards serves as the basis of education development strategy based on the result of evaluation the implementation of education. The evaluation is implemented through accreditation and national exam. The objective of this study is to analyze the score of computer-based national exam based on accreditation scores per items of instrument by applying multiclass random forest classification modeling. The research used Computer-Based National Exam data in 2018 and accreditation data from the year of 2017 and 2018. This study employed random forest for multiclass classification. The results showed that, based on the evaluation model, classification accuration value in multiclass random forest was 83.49%. In addition, this model produces important variables in classifying the average value of computer-based national examination, i.e., items laboratory conditions (x71, x68, x69, x67), electrical installation (x62), infrastructure (x64), canteen (x83), laboratory (x55), special service officers (x56), certified teachers (x39), library staff (x54), head of administration (x51), literacy activities for students (x33), use of textbooks (x14), and community/partner collaboration in education management (x96). Based on the indicators of important variables, National Education Standards that have important role are facility and infrastructure standards, educator and educational staff standards, and graduate competence standards. Therefore, improving the quality of education can be done by improving school facilities, the competency of teacher and education staff, and graduate competency.

The Program for International Student Assessment (PISA), becomes one of the references or indicators used to assess the development of students' knowledge and skills in each member country of the Organization for Economic Cooperation and Development (OECD). The results of the PISA survey in 2018 placed Indonesia in the bottom 10, indicating that the implementation of the national education system has not been successful. This underlies the need for a more in-depth study of the factors that influence PISA data scores not only statistically qualitatively but also quantitatively which is still very rarely done. The data structure of the PISA survey results is complex, which involves multicollinearity, multivariate response variables, and random effects. Thus, it requires an appropriate statistical analysis method such as the multivariate mixed linear regression (MLMM) model. In this study, secondary data from the results of the 2018 PISA survey with Indonesian students as the smallest unit of observation were used as sample. School is used as an intercept random effect which is assumed to be normally distributed. Multicollinearity is overcome by selecting independent variables based on AIC and BIC values. Estimation of variance and random effect parameters was performed using the restricted maximum likelihood (REML) method. Based on the estimator of the variance of random effects for the response variables of mathematics, science, and reading literacy, it was obtained 1548.12, 1359.39, and 1082.48, respectively, which explains the significant effect of each school as a random effect on the three response variables.

No abstract