PERBANDINGAN KLASIFIKASI HASIL INDEKS KEPUASAN MASYARAKAT TERHADAP UPT BALAI LATIHAN KERJA SURABAYA MENGGUNAKAN ALGORITMA DECISION TREE, RANDOM FOREST, K-NEAREST NEIGHBOR

Abtract – The evaluation of the results of the community statisfaction index towards the UPT Job Training Center Surabaya which was carried out by giving questionnaires to the community and training participants at UPT Job Training Center Surabaya, then the questionnaires file was submited to the Development dan Marketing Staff. However, the results of the questionaires which should be given to the Head Of Surabaya Job Training Center as an evaluation material for the UPT Job Training Center Surabaya is often just neglected and not processed further by the UPT Job Training Center Surabaya. So that the evaluation of the questionaires, especially on infrastructure through a sheet of paper, can be considered ineffective and not on target. To overcome the problems above, the author provides an alternative by utilizing the knwoledge gained during lectures using data mining techniques to predict community satisfation using a comparison method of 3 algorithms, namely Decision Tree, Random Forest, K-Nearest Neighbor. Classification is a technique in data mining to classify data based on data againts sample data. In study, the authors compare 3 classification techniques to classify community satisfaction responses in the dataset of UPT JobTraining Center Surabaya. The classification process begins with preprocessing the data ti remove mising values and select features in the dataset. At the evaluation stage, confusion matrix and K-fold cross validation techniques are used. After testing, it was found the classification results showed the best accuracy obtained by the Random Forest model obtaining an accuracy of 1.00, Decision Tree obtaining an accuracy of 0.91, and K-Nearest Neighbor 0.94.

Keywords : Comparison, Classification, data mining, Decision Tree, Random Forest, K-Nearest Neighbor

Abstrak – Pelaksanaan evaluasi hasil indeks kepuasan masyarakat terhadap UPT Balai Latihan Kerja Surabaya yang dilakukan dengan cara memberikan kuisioner kepada masyarakat dan perserta pelatihan yang ada di UPT Balai Latihan Kerja Surabaya, kemudian berkas kuisioner diserahkan kepada pihak Staff Pengembangan dan Pemasaran.. Akan tetapi hasil kuisioner yang seharusnya diberikan kepada kepala UPT Balai Latihan Kerja Surabaya sebagai bahan evaluasi UPT Balai Latihan Kerja Surabaya seringkali hanya terbengkalai dan tidak diproses lebih lanjut oleh UPT Balai Latihan Kerja Surabaya. Sehingga evaluasi kuisioner  khususnya pada sarana prasarana melalui selembaran kertas bisa dianggap tidak efektif dan tidak sesuai sasaran. Untuk mengatasi permasalah  diatas penulis memberikan alternatif dengan cara memanfaatkan ilmu yang diperoleh selama perkuliahan dengan menggunakan teknik data mining untuk memprediksi kepuasan masyarakat  menggunakan metode perbandingan 3 algoritma yaitu Decision Tree, Random Forest, K-Nearest Neighbor.  Klasifikasi merupakan Teknik dalam data mining untuk mengklompokkan data berdasarkan data terhadap data sampel. Pada penelitian ini, penulis melakukan perbandingan 3 teknik klasifikasi untuk mengklasifikasikan respon kepuasan masyarakat pada dataset UPT Balai Latihan Kerja Surabaya. Perbandingan teknik klasifikasi ini dilakukan untuk mengetahui model dalam teknik klasifikasi mana yang paling efektif untuk mengklasifikasikan target pada dataset UPT Balai Latihan Kerja Surabaya. Proses klasifikasi diawali dengan preprocessing data untuk melakukan penghilangan missing value dan pemilihan fitur pada dataset. Pada tahap evaluasi digunakkan teknik confusion matrix dan Kfold cross validation. Setelah dilakukan pengujian, didapatkan bahwa hasil klasifikasi menunjukan akurasi terbaik diperoleh oleh model Random Forest memperoleh accuracy 1.00, Decision Tree memperoleh accuracy 0.91 , dan K-Nearest Neighbor 0.94.

Kata Kunci : Perbandingan, Klasifikasi, data mining, Decision Tree, Random Forest, K-Nearest Neighbor

Abtract – The evaluation of the results of the community statisfaction index towards the UPT Job Training Center Surabaya which was carried out by giving questionnaires to the community and training participants at UPT Job Training Center Surabaya, then the questionnaires file was submited to the Development dan Marketing Staff. However, the results of the questionaires which should be given to the Head Of Surabaya Job Training Center as an evaluation material for the UPT Job Training Center Surabaya is often just neglected and not processed further by the UPT Job Training Center Surabaya. So that the evaluation of the questionaires, especially on infrastructure through a sheet of paper, can be considered ineffective and not on target. To overcome the problems above, the author provides an alternative by utilizing the knwoledge gained during lectures using data mining techniques to predict community satisfation using a comparison method of 3 algorithms, namely Decision Tree, Random Forest, K-Nearest Neighbor. Classification is a technique in data mining to classify data based on data againts sample data. In study, the authors compare 3 classification techniques to classify community satisfaction responses in the dataset of UPT JobTraining Center Surabaya. The classification process begins with preprocessing the data ti remove mising values and select features in the dataset. At the evaluation stage, confusion matrix and K-fold cross validation techniques are used. After testing, it was found the classification results showed the best accuracy obtained by the Random Forest model obtaining an accuracy of 1.00, Decision Tree obtaining an accuracy of 0.91, and K-Nearest Neighbor 0.94.

Keywords : Comparison, Classification, data mining, Decision Tree, Random Forest, K-Nearest Neighbor