Articles

Found 5 Documents
Search
Journal : PIKSEL (Penelitian Ilmu Komputer Sistem Embedded dan Logic)

PENERAPAN ALGORITMA K-NEAREST NEIGHBOR UNTUK PENENTUAN RESIKO KREDIT KEPEMILIKAN KENDARAAN BEMOTOR Leidiyana, Henny
PIKSEL (Penelitian Ilmu Komputer Sistem Embedded dan Logic) Vol 1, No 1 (2013): PIKSEL (Penelitian Ilmu Komputer Sistem Embedded dan Logic
Publisher : PIKSEL (Penelitian Ilmu Komputer Sistem Embedded dan Logic)

Show Abstract | Original Source | Check in Google Scholar

Abstract

Sejalan dengan pertumbuhan bisnis, kredit merupakah masalah yang menarik untuk diteliti. Beberapa riset bidang komputer untuk mengurangi resiko kredit telah banyak dilakukan dalam rangka menghindarai kehancuran suatu perusahaan pembiayaan.  Paper ini membahas algoritma k-Nearest Neighbor (kNN) yang diterapkan pada data konsumen yang menggunakan jasa keuangan kredit kendaraan bermotor. Hasil testing untuk mengukur performa algoritma ini menggunakan metode Cross Validation, Confusion Matrix dan kurva ROC dan menghasilkan akurasi dan nilai AUC berturut-turut 81,46 % dan 0,984. Karena nilai AUC berada dalam rentang 0,9 sampai 1,0 maka metode tersebut masuk dalam kategori sangat baik (excellent).  Kata kunci : K-Nearest Neighbor, Cross Validation, Confusion matrix, ROCIn line with the growth and business development, credit issues remain to be studied and revealed interesting. Some of the research field of computers has done much to reduce the credit risk of causing harm to the company. In this study, k-Nearest Neighbor (kNN) algorithm is applied to the data of consumers who have good credit financing motorcycle that consumers are troubled or not. From the test results to measure the performance of the algorithms using the test method Cross Validation, Confusion Matrix and ROC curves, it is known that the accuracy value of  81.46% and AUC values of 0.984. This methodes is include excellent classification because the AUC value between 0.90-1.00.  Keywords: K-Nearest Neighbor, Cross Validation, Confusion matrix, ROC
PENENTUAN KLASIFIKASI MUTU FISIK BERAS DENGAN METODE NAÏVE BAYES Leidiyana, Henny
PIKSEL (Penelitian Ilmu Komputer Sistem Embedded dan Logic) Vol 1, No 2 (2013): PIKSEL (Penelitian Ilmu Komputer Sistem Embedded dan Logic)
Publisher : PIKSEL (Penelitian Ilmu Komputer Sistem Embedded dan Logic)

Show Abstract | Original Source | Check in Google Scholar

Abstract

Rice is the staple food of the people of Indonesia in particular. One factor that made the peoples choice in selecting rice is quality. Writing is about the determination of the classification of rice quality is acceptable and not based on quality components that have been set by the ISO.A total of 1161 data is used as training data. The data obtained consists of 6 attributes, namely moisture content, milling degree, broken grains, grain groats, because the other is a predictor attributes, while the condition of a label attribute (class). Training data is cleaned and then made a physical model of the determination of quality grade rice using Naïve  Bayes method. The resulting model was tested using the method of Cross Validation and the ROC curve. From the test results obtained by the results of model accuracy by 92.56% and AUC values for 0989, this means that the resulting model is classified as very goodKeywords : Naïve  Bayes , Cross Validation, Confusion matrix, ROC curveBeras merupakan makanan pokok masyarakat Indonesia khususnya. Salah satu faktor yang dijadikan pilihan masyarakat dalam memilih beras adalah mutunya. Penulisan ini membahas tentang penentuan klasifikasi mutu beras yang dapat diterima dan tidak berdasarkan komponen mutu yang telah ditetapkan oleh SNI. Sebanyak 1161 data digunakan sebagai data training. Data yang didapat terdiri dari 6 atribut, yaitu kadar air, derajat sosoh, butir patah, butir menir, sebab lain merupakan atribut prediktor, sedangkan kondisi merupakan atribut label (kelas). Data training dibersihkan kemudian dibuat model penentuan kelas mutu fisik beras menggunakan metode Naïve  Bayes . Model yang dihasilkan diuji menggunakan metode Cross Validation dan Kurva ROC.  Dari hasil pengujian diperoleh hasil akurasi model sebesar 92,56% dan nilai AUC sebesar 0.989, ini berarti bahwa model yang dihasilkan termasuk klasifikasi sangat baik.Kata Kunci : Naïve  Bayes , Cross Validation, Confusion matrix, Kurva ROC
PREDIKSI PINJAMAN KREDIT DENGAN SUPPORT VECTOR MACHINE DAN K-NEAREST NEIGHBORS PADA KOPERASI SERBA USAHA Iriadi, Nandang; Leidiyana, Henny
PIKSEL (Penelitian Ilmu Komputer Sistem Embedded dan Logic) Vol 1, No 2 (2013): PIKSEL (Penelitian Ilmu Komputer Sistem Embedded dan Logic)
Publisher : PIKSEL (Penelitian Ilmu Komputer Sistem Embedded dan Logic)

Show Abstract | Original Source | Check in Google Scholar

Abstract

Cooperatives as a form of organization that are important in promoting economic growth . Cooperatives be an alternative for people to get funds in an effort to improve their quality of life , day-to- day needs and develop the business . No doubt , lend funds to member cooperatives will surely emerge problems , such as members of the borrower paying the overdue installment of funds , misuse of funds for other purposes , the customer fails to develop its business so as to result in cooperative funds do not flow or it can lead to bad credit . In this research will be carried out loans prediction using data mining classification Support Vector Machine and k - Nearest Neighbors were then conducted a comparison of both methods . From the test results to measure the performance of both methods using cross validation , confusion matrix and ROC curves is known that Support Vector Machine has an accuracy value of 92.67 % followed by k -Nearest Neighbors, which has a value of 88.67 % accuracy . Thus the Support Vector Machine method is included in Verry Good Clasification because it has the accuracy of 92.67 % .Keywords: comparative, Support Vector Machines, k-Nearest Neighbors, Credit Analysis Cooperatives as a form of organization that are important in promoting economic growth . Cooperatives be an alternative for people to get funds in an effort to improve their quality of life , day-to- day needs and develop the business . No doubt , lend funds to member cooperatives will surely emerge problems , such as members of the borrower paying the overdue installment of funds , misuse of funds for other purposes , the customer fails to develop its business so as to result in cooperative funds do not flow or it can lead to bad credit . In this research will be carried out loans prediction using data mining classification Support Vector Machine and k - Nearest Neighbors were then conducted a comparison of both methods . From the test results to measure the performance of both methods using cross validation , confusion matrix and ROC curves is known that Support Vector Machine has an accuracy value of 92.67 % followed by k -Nearest Neighbors, which has a value of 88.67 % accuracy . Thus the Support Vector Machine method is included in Verry Good Clasification because it has the accuracy of 92.67 % .Keywords: comparative, Support Vector Machines, k-Nearest Neighbors, Credit Analysis
PENENTUAN KLASIFIKASI MUTU FISIK BERAS DENGAN METODE NAÏVE BAYES Leidiyana, Henny
PIKSEL (Penelitian Ilmu Komputer Sistem Embedded dan Logic) Vol 1 No 2 (2013): September 2013
Publisher : LPPM Universitas Islam 45 Bekasi

Show Abstract | Original Source | Check in Google Scholar

Abstract

ABSTRACTRice is the staple food of the people of Indonesia in particular. One factor that made the people's choice inselecting rice is quality. Writing is about the determination of the classification of rice quality is acceptableand not based on quality components that have been set by the ISO.A total of 1161 data is used as trainingdata. The data obtained consists of 6 attributes, namely moisture content, milling degree, broken grains,grain groats, because the other is a predictor attributes, while the condition of a label attribute(class). Training data is cleaned and then made a physical model of the determination of quality grade riceusing Naïve Bayes method. The resulting model was tested using the method of Cross Validation and theROC curve. From the test results obtained by the results of model accuracy by 92.56% and AUC values for0989, this means that the resulting model is classified as very good Keywords : Naïve Bayes, Cross Validation, Confusion matrix, ROC curve ABSTRAKBeras merupakan makanan pokok masyarakat Indonesia khususnya. Salah satu faktor yang dijadikanpilihan masyarakat dalam memilih beras adalah mutunya. Penulisan ini membahas tentang penentuanklasifikasi mutu beras yang dapat diterima dan tidak berdasarkan komponen mutu yang telah ditetapkanoleh SNI. Sebanyak 1161 data digunakan sebagai data training. Data yang didapat terdiri dari 6 atribut,yaitu kadar air, derajat sosoh, butir patah, butir menir, sebab lain merupakan atribut prediktor, sedangkankondisi merupakan atribut label (kelas). Data training dibersihkan kemudian dibuat model penentuan kelasmutu fisik beras menggunakan metode Naïve Bayes. Model yang dihasilkan diuji menggunakan metodeCross Validation dan Kurva ROC. Dari hasil pengujian diperoleh hasil akurasi model sebesar 92,56% dannilai AUC sebesar 0.989, ini berarti bahwa model yang dihasilkan termasuk klasifikasi sangat baik. Kata Kunci : Naïve Bayes, Cross Validation, Confusion matrix, Kurva ROC
PREDIKSI PINJAMAN KREDIT DENGAN SUPPORT VECTOR MACHINE DAN K-NEAREST NEIGHBORS PADA KOPERASI SERBA USAHA Iriadi, Nandang; Leidiyana, Henny
PIKSEL (Penelitian Ilmu Komputer Sistem Embedded dan Logic) Vol 1 No 2 (2013): September 2013
Publisher : LPPM Universitas Islam 45 Bekasi

Show Abstract | Original Source | Check in Google Scholar | Full PDF (248.691 KB)

Abstract

BSTRACTCooperatives as a form of organization that are important in promoting economic growth . Cooperatives bean alternative for people to get funds in an effort to improve their quality of life , day-to- day needs anddevelop the business . No doubt , lend funds to member cooperatives will surely emerge problems , such asmembers of the borrower paying the overdue installment of funds , misuse of funds for other purposes , thecustomer fails to develop its business so as to result in cooperative funds do not flow or it can lead to badcredit . In this research will be carried out loans prediction using data mining classification Support VectorMachine and k - Nearest Neighbors were then conducted a comparison of both methods . From the testresults to measure the performance of both methods using cross validation , confusion matrix and ROCcurves is known that Support Vector Machine has an accuracy value of 92.67 % followed by k -NearestNeighbors, which has a value of 88.67 % accuracy . Thus the Support Vector Machine method is includedin Verry Good Clasification because it has the accuracy of 92.67 % . Keywords: comparative, Support Vector Machines, k-Nearest Neighbors, Credit Analysis ABSTRAKKoperasi sebagai salah satu bentuk organisasi yang penting dalam meningkatkan pertumbuhan ekonomi.Koperasi simpan pinjam menjadi salah satu alternatif bagi masyarakat untuk mendapatkan dana dalamupaya memperbaiki taraf kehidupan, pemenuhan kebutuhan sehari-hari dan mengembangkan usaha.Tidakdipungkiri, memberikan pinjaman dana kepada anggota koperasi pasti akan muncul permasalahanpermasalahan, seperti anggota peminjam terlambat membayarkan cicilan dana, penyalahgunaan dana untukkeperluan lain, nasabah gagal mengembangkan usahanya sehingga dapat mengakibatkan dana di koperasitidak mengalir atau dapat mengakibatkan kredit macet. Dalam penelitian ini akan dilakukan prediksipinjaman kredit dengan menggunakan metode klasifikasi data mining Support Vector Machine dan kNearest Neighbor syang kemudian dilakukan komparasi kedua metode tersebut. Dari hasil pengujiandengan mengukur kinerja kedua metode tersebut menggunakan cross validation, confusion matrix dankurva ROC diketahui bahwa Support Vector Machine memiliki nilai akurasi 92.67% diikuti oleh k-NearestNeighbors yang memiliki nilai akurasi 88,67%. Dengan demikian Metode Support Vector Machine tersebuttermasuk dalam Verry Good Clasification karena memiliki nilai akurasinya sebesar 92.67%. Kata kunci: komparasi,Support Vector Machine,k-Nearest Neighbors ,Analisa Kredit