The purpose of this research is to see how much open unemployment rate according to the highest education completed in the country of Indonesia for subsequent years through predictions used on the basis of existing data, which later as input for the government so that the government can make better policies to suppress the unemployment rate. This research uses artificial neural network application using a combination of Levenberg-Marquardt Algorithm with bipolar sigmoid function. Open unemployment data according to the highest education is sourced from the National Labor Force Survey of the Republic of Indonesia, 2013-2017 in each semester. The data processing consists of two stages where the first phase of pattern recognition and the second stage is predicted. Pattern recognition and prediction use different data from the same process that uses data training and data testing. Data Training year 2013-2015 with a target of 2016, while data testing year 2014-2016 with the target year 2017. Architectural model used there are five, among others 6-2-5-2, 6-5-6-2, 6- 5-8-2, 6-5-10-2 and 6-8-12-2. From the 5 models, it can be concluded that the best model is 6-5-10-2 with the epoch of 13 iterations, MSE in February 0.0109696004, MSE in August 0.0233797200. While the accuracy rate in February and August is the same, that is equal to 88%.

The purpose of this study can predict the feasibility of the location of the development of clean water sources in the Tirta Lihou PDAM using the Naive Bayes algorithm. With the increasing number of MBR (Low-Income Communities) that enter each year in each region, the Tirta Lihou PDAM plans to find alternative springs solutions for several production units so that they can meet the needs of the community. In determining the appropriate alternative sources of clean water in several production units, the datamining method is used. By using data mining techniques specifically classification using the Naive Bayes algorithm, predictions can be made on the feasibility of the location of the construction of clean water sources based on existing data. Naive bayes is a simple probabilistic prediction technique based on the Bayes theorem with a strong assumption of independence (dependence). Based on the results of calculations using algoritma naive bayes, the clasification results from 19 alternatives used, where there are 8 feasible classes and 11 classes are not feasible with the number of accuracy obtained at 78,95%. From the results obtained, it is expected that this research can help the PDAM Tirta Lihou in determining the location that is feasible to develop water sources so that it can meet the needs of the community. This research is also expected to be a reference for further researchers relating to the user algorithm used.

Based on data on the results of oil palm production in PTPN IV Marihat displays several locations with fruit yields that vary in number. For this reason, grouping of potential fruit-producing locations is needed to know which locations produce large or small numbers of palm fruit. The production sharing is usually done based on the location or block of harvesting oil palm fruit. Therefore, a method is needed to facilitate the grouping of fruit producing locations. With the K-Means clustering approach, the division of location groups can be done based on harvested area (Ha), production realization (kg) and harvest year. In this research, clustering of potential fruit-producing areas was carried out using the K-Means algorithm. By using K-Means aims to facilitate the grouping of a block with a lot of fruit production, and low. The result of this research is that C1 (highest) is 14 Harvest Block data, and C2 (lowest) is 11 Harvest Block data.

This study aims to classify the level of understanding of students at STIKOM Tunas Bangsa Pematangsiantar. STIKOM Tunas Bangsa is one of the private universities in North Sumatra that is engaged in the field of computer science. In carrying out lecture activities, students are required to understand each lecture material provided by the lecturer. There are several things that can affect the level of understanding of students in receiving lecture material. The data source was obtained from the results of the fifth semester and seven student questionnaires in the STIKOM Tunas Bangsa Information System department. Attributes used are as many as five, namely communication, learning atmosphere, learning media, appearance and how to teach. The method used in the research is C4.5 Algorithm and assisted by RapidMiner software to make decision trees. From the results of the study, there were 14 rules for classification in determining the level of understanding with 9 rules, the best status and 5 rules did not understand. C4.5 algorithm can be used in the case of determining the level of understanding of students at STIKOM Tunas Bangsa with an accuracy rate of 87.10%. With this analysis it is expected to be a motivation for students to be able to understand the course well.

The Livestock Sector is one of the most promising agribusiness sectors. The selection of the right type of cow is the duty of cattle farmers to get cows with good quality and greatly affect the success in raising cattle. Decision support system is a system that is able to provide a picture of the decision of the existing situation. And this system is applied to cattle farms to determine the best type of cow for the smooth running of the farm business. The method used is SMART. The samples used were 6 types of cattle and the assessment criteria in the selection of cattle were: Origin, Price, Age, Weight, and Size. The results of this study were the selection of Lemosin cattle as the best type of cattle for beef cattle farming. SPK in Cattle Purchases for Breeding is only a recommendation to the farmer, for the next process to be returned to the farmer.