To solve the data redundancy problem that occurs in the STT - PLN library automation system, namely the Senayan and Cendana systems, currently a web service has been designed to combine the two automation systems in the library so that data redundancy from the two systems will not occur. This system was created in order to prepare a database for the implementation of a new automation system for the STT - PLN library, namely the Akasia system. This web service application can migrate and merge data from the Senayan and Cendana databases to the Akasia database. This system also supports the creation of additional fields and tables for fields and tables that were not yet available in the initial structure of the acacia database. The application of the REST architecture here is to test the ability of the REST architecture in handling data migration and merging with large amounts of data. The test is carried out on a local host and through the existing LAN network in the IT - PLN building. There are two types of testing, namely speed testing and validation testing. The speed test is done by calculating the estimated time spent when transferring data with the amount of data ranging from 4000, 6000, 8000, 10000.The result is that using the REST architecture in migrating data can shorten the migration time because the data transferred from the source database is in JSON format. which has a smaller size than the data in SQL format. The data in JSON format is then converted into an array by the RESTful API and then entered into the acacia database. So that the migration process has a short time because the process of transferring data from the source database server to the RESTful API server is in JSON format. While the results of data validation testing using RMSE show a very small number, namely 0. From these results it can be stated that the acacia database from the results of the localhost migration already has sufficient data accuracy to be used as a new system database.

The prediction of the number of courses is done by the department before making a schedule for each course. In practice, the number of classes in each course has a different number and there is often an opening or closing class when compiling a KRS due to the number of classes that are not in accordance with the number of students. A system is needed to produce a number of classes so that it can reduce the number of opening classes because the demand for a higher number of classes is in the class because of the interest in a class that will be opened. Fuzzy methods are used to predict students who will repeat the course based on student force and value variables. The K-Means method is used to classify the subjects with the number of students converted into 2 groups based on the number of students who have been taken and the number of students who repeat a number of subjects. The two methods used are implemented in the application system to predict the number of classes. The results of the fuzzy and K-method processes mean the output of the application predictions the number of classes.