Analisa Nilai Lamda Model Jarak Minkowsky Untuk Penentuan Jurusan SMA (Studi Kasus di SMA Negeri 2 Tualang)

Abstract

SMA Negeri 2 (SMAN 2) is located in Tualang. So far the data report student majors only stored in a database as a final report. Data from the report of the majors could be used as guidelines to determine the students' decision majors for the following year. To take advantage of the data stored in that particular database, we can use data mining disciplines. The method used to make the determination of students majoring done by using Nearest K-Nearest Neighbor (K-NN) algorithm. On the other hand, the method for calculating the distance between the data used models Minkowsky distance with a value of lambda (λ) as a parameter. Lambda values that were analyzed were lambda 1, 2 and 3. Lambda with the value of 1 can generate increasing accuracy in the 11th experiment or with a large amount of data equal to 276 data. Lambda 2 will produce increasing accuracy by the 16th experiment or with the number of training data equal to 356 data while lambda 3 can also produce accuracy continuously increasing by the 11th experiement or with the amount of training data equal to 276 data. The accuracy of the lambda value of 1 is better than lambda 2 and lambda 3. This was proven in 25 experiments at lambda 1 which produces the highest accuracy value for 20 times.Keywords — Classification, Data Mining, K-Nearest Neighbor, Lamda (λ), Minkowsky.