Indonesia has a capital city which is one of the many big cities in the world called Jakarta. Jakarta's role in the dynamics that occur in Indonesia is very central because it functions as a political and government center, and is a business and economic center that drives the economy. Recently the discourse of the government to relocate the capital city has invited various reactions from the community. Therefore, in this study, sentiment analysis of the relocation of the capital city was carried out. The analysis was performed by doing a classification to describe the public sentiment sourced from twitter data, the data is classified into 2 classes, namely positive and negative sentiments. The algorithms used in this study include Naïve Bayes classifier, logistic regression, support vector machine, and K-nearest neighbor. The results of the performance evaluation algorithm showed that support vector machine outperformed as compared to 3 algorithms with the results of Accuracy, Precision, Recall, and F-measure are 97.72%, 96.01%, 99.18%, and 97.57%, respectively. Sentiment analysis of the discourse of relocation of the capital city is expected to provide an overview to the government of public opinion from the point of view of data coming from social media.

No abstract

Forest and land fires are disasters that often occur in Indonesia. In 2007, 2012 and 2015 forest fires that occurred in Sumatra and Kalimantan attracted global attention because they brought smog pollution to neighboring countries. One of the regions that has the highest fire hotspots is West Kalimantan Province. Forest and land fires have an impact on health, especially on the communities around the scene, as well as on the economic and social aspects. This must be overcome, one of them is by knowing the location of the area of ??fire and can analyze the causes of forest and land fires. With the impact caused by forest and land fires, the purpose of this study is to apply the clustering method using the k-means algorithm to be able to determine the hotspot prone areas in West Kalimantan Province. And evaluate the results of the cluster that has been obtained from the clustering method using the k-means algorithm. Data mining is a suitable method to be able to find out information on hotspot areas. The data mining method used is clustering because this method can process hotspot data into information that can inform areas prone to hotspots. This clustering uses k-means algorithm which is grouping data based on similar characteristics. The hotspots data obtained are grouped into 3 clusters with the results obtained for cluster 0 as many as 284 hotspots including hazardous areas, 215 hotspots including non-prone areas and 129 points that belong to very vulnerable areas. Then the clustering results were evaluated using the Davies-Bouldin Index (DBI) method with a value of 3.112 which indicates that the clustering results of 3 clusters were not optimal.

At present, the use of ICT is growing very rapidly. This has led to changes in processes, functions, and policies in various sectors, including the public service sector managed by the government. e-Government is a new mechanism between the government and the community and stakeholders, which involves the implementation of information technology, and aims to improve the quality (quality) of public services. Bandung City is one of the cities that is very intensive in developing the use of ICT in implementing e-Government. The focus of the city government of Bandung is the Government to Citizen (G2C) application model. The application that is still lacking in use is the e-punten application. One important factor for the success of e-Government services is the acceptance and willingness of people to use e-Government services. E-punten application services provided by the Bandung city government will not run perfectly if no people are using it. To assess what factors influence the use of e-punten applications in the city of Bandung, the UTAUT model is used. To analyze the factors that influence the acceptance of e-punten applications SEM analysis is used. In this study, the PLS-SEM approach is used to solve multiple regression when specific problems occur in the data, such as the small sample size of the study. The PLS evaluation is carried out by evaluating the measurement model and the structural model that best suits the UTAUT model. Factors that influence the use of e-punten applications are effort expectancy for behavioral intention, facilitating conditions for use behavior, and behavioral intention for use behavior. The factors that have the most influence are performance expectancy and effort expectancy.

Dalam data nyata, ada banyak situasi di mana jumlah instance di satu class jauh lebih sedikit daripada jumlah instance di class lain. Keadaan ini disebut sebagai masalah dataset tidak seimbang (imbalance class). Imbasnya kinerja klasifikasi biasanya menurun di beberapa aplikasi data mining. Pada penelitian ini, diidentifkasi bahwa dataset performansi rating iklan TV yang digunakan memiliki permasalahan imbalance class yang sangat besar dimana instance yang memiliki nilai rating tinggi, jauh lebih sedikit dibandingkan instance yang memiliki nilai rating kecil dan menengah. Sehingga diperlukan metode over-sampling untuk mengatasi permasalahan imbalance class tersebut. Metode yang dapat digunakan adalah Synthetic Minority Over-sampling Technique (SMOTE). Untuk memvalidasi keefektifan model yang diusulkan, dilakukan dua skenario eksperimental yaitu: pertama algoritma ANN langsung digunakan untuk pemodelan tanpa mempertimbangkan ketidakseimbangan kelas, dan kedua dilakukan over-sampling SMOTE untuk meningkatkan jumlah dataset agar mencapai dataset yang seimbang. Hasil eksperimen menunjukkan bahwa performansi ANN+SMOTE mencapai akurasi sebesar 87.06% dibandingkan ANN yang hanya sebesar 86.35%. Penerapan Teknik SMOTE terbukti dapat mengatasi masalah ketidakseimbangan data dan mendapatkan hasil klasifikasi yang lebih baik.