1,720,998 research outputs found
Klasifikasi Debitur Kartu Kredit dengan Pemilihan Fitur Menggunakan Voting Feature Intervals 5
No full textProvision of credit cards for customers is one of the ways to obtain profit in banking activities which cause risks of losses if the customer frequently delinquent the payments. Therefore, it is important to know the banking profile of the customer who will apply for a credit card. The banking profile data is used as input for Voting Feature Intervals 5 (VFI5) algorithm in the development of classification models that aim to classify potential debtor based on the payment status of the debtor. The debtor data used in this research is categorized as imbalanced data, hence it is necessary to have other performance measures beside accuracy; in this research we also used recall and precision. The input data consist of 14 features, however each features has different significance in classifying debtor. Therefore a feature selection process is conducted before the development of the model. The feature selection is conducted using two approaches: feature selection based on the accuracy of each feature and stepwise feature selection. The former method provides the better accuracy of 67.74%, and the values of recall and precision for the class of bad debtor are 46.88% and 24.69%, respectively
Transformasi Koordinat Menggunakan Jaringan Syaraf Tiruan Propagasi Balik Resilient
No full textGenuk datum, which is based on Bessel 1841 ellipsoid model, is a local datum used in Indonesia. On the other hand, World Geodetic System 1984 (WGS-84), a datum used in GPS measurement, is commonly used by the rest of the world. Therefore in order, to fully utilize WGS- 84, Genuk datum has to be transformed to WGS-84 datum. A model is needed to transform coordinates from Genuk datum into WGS-84 datum. This research discusses coordinate transformation using similarity transformation (4-parameter) and Resilient Back-propagation Neural Network. Similarity transformation is a traditional method for coordinate transformation. The Resilient Back-propagation Neural Network provides a new technology for coordinate transformation. Coordinate transformation in this research is conducted on West Java coordinate data. The data are split into two parts: a third are used as testing data and the rest are used as training data. The test results show that the coordinate transformation using Resilient Backpropagation Neural Network can be used as an alternative model to coordinate transform
Identifikasi Varietas Durian Berdasarkan Tekstur Daun Menggunakan K-Nearest Neighbor Dengan Ciri Statistical Textures
No full textDurian (Durio zibethinus Murray) adalah salah satu komoditas yang
memiliki nilai ekonomi yang tinggi di Indonesia. Nilai ekonomi durian
dipengaruhi oleh keunggulan yang dimiliki setiap varietasnya. Oleh sebab itu
identifikasi tanaman durian merupakan hal yang sangat penting. Pada penelitian
ini, identifikasi dilakukan berdasarkan citra daun dari empat varietas durian.
Metode klasifikasi yang digunakan adalah K-Nearest Neighbor dengan ekstraksi
ciri statistical textures. Penelitian ini menghasilkan akurasi tertinggi sebesar
67,5% menggunakan 6 ciri statistical textures. Pada penelitian ini paling sulit
teridentifikasi varietas Sukun. Varietas Sukun mempunyai kemiripan ke varietas
Bakul dan Matahari
Identifikasi Daun Shorea menggunakan KNN dengan Ekstraksi Fitur 2DPCA
No full textShorea is a kind of meranti species that have high economic value. Shorea belongs to Dipterocarpaceae family which has 194 species that usually grow in tropical area. Shorea is difficult to be identified due to their similarity. The method used in this research is 2 Dimensional Principal Component Analysis as feature extraction and K-Nearest Neighbour as classification technique. This research has four trials that were divided into R, G, B, and grayscale components. The best average accuracy obtained was 75% on a G component with 85% contribution of eigen values
Identifikasi Tanaman Durian Berdasarkan Citra Daun Menggunakan LVQ dan Ekstraksi Tekstur Discrete Wavelet Transform
No full textDurian (Durio zibethinus Murray) adalah salah satu komoditas yang memiliki nilai ekonomi tinggi. Nilai ekonomi durian dipengaruhi oleh keunggulan yang dimiliki setiap varietasnya. Oleh sebab itu identifikasi tanaman durian merupakan hal yang sangat penting. Identifikasi dilakukan dengan menggunakan citra daun dari lima tanaman varietas durian, serta metode klasifikasi Learning Vector Quantization (LVQ) dan ekstraksi tekstur Discrete Wavelet Transform (DWT). Penelitian ini menghasilkan akurasi tertinggi sebesar 76% pada DWT famili Haar level 6
Prediksi Indeks Prestasi Mahasiswa Menggunakan Algoritma VFI5 (Studi Kasus Mahasiswa Program Mayor Minor Departemen Ilmu Komputer IPB)
No full textFeature Voting Intervals 5 describes a concept by performing classification on each feature separately. VFI5 is a non-incremental algorithm in which all training instances are processed simultaneously. Two steps of VFI5 algotirhms are training and predicting steps. Training is used to explain the character and relationships between features, and predicting is used to test the pattern resulted from training step to get accuration value. This study applied VFI5 algorithm to predict GPA value of the following academic year based on feature which is value of current academic course. First step training was conducted to explain the distribution class of student’s GPA and the ability of college students, while the second step predicting can reveal opportunities to learn and know the GPA value that can be obtained by students in the following academic year. The set of data had been used for this research, the first is first year student data that predicted to second year student GPA from generation 2005/2006, second is second year student data that predicted to third year student GPA from generation 2005/2006, and the last is first year student data that predicted to second year data GPA from generation 2006/2007. In the training step, features on the first data which value equals to a given class are Ekonomi Umum and Pengantar Matematika, features on the second data which value equals to a given class are comonly all features, and the last data which value equals to a given class are Ekonomi Umum, Agama, Bahasa Indonesia, PIP, Pengantar Matematika and Pengantar Kewirausahaan. In predicting step, process had been done with and without GPA feature. As the result, we got accuration value in the first data with GPA feature is 45.68 %, and without GPA feature is 46.91 %. In the second data with GPA feature got value 49.56 % and without GPA feature got value 50.80 %. In the last data with GPA feature got value 60.28 % and without GPA feature got value 51.80 %. The result showed that there was a declining accuration as compared to the previous studies due to differences in the testing data and type of prediction
Pengenalan Beras Campuran Menggunakan Transformasi Wavelet dan Probabilistic Neural Network.
No full textPenelitian identifikasi varietas beras menggunakan pengolahan citra digital menjadi penting karena dapat digunakan sebagai alternatif dalam mengidentifikasi varietas beras. Pada penelitian ini, identifikasi varietas beras dilakukan dengan menggunakan ekstraksi ciri transformasi wavelet dan pengklasifikasian menggunakan metode Probabilistic Neural Network. Identifikasi citra beras dilakukan dengan 3 percobaan yaitu identifikasi varietas beras tunggal, beras campuran, dan gabungan dari keduanya. Varietas beras yang digunakan adalah beras jagung, beras ketan putih, beras pandan wangi, dan beras rojo lele. Akurasi terbaik sebesar 90% diperoleh pada dekomposisi Wavelet level 6 menggunakan data citra gabungan dari varietas beras tunggal dan beras campuran
Identifikasi Mangrove Berbasis Citra Daun Menggunakan KNN dengan Ekstraksi Tekstur Wavelet
No full textMangroves are plants that live in tidal area. Mangroves have many benefits such as preventing abrasion and becoming medicinal plants. Mangroves identification is difficult because of their various species and simililarities between species. This research developed a system to identify Mangrove using Discrete Wavelet Transform and K-Nearest Neighbour classification based on mangrove leaf image. The best accuracy in this research was 88.75%, obtained at Discrete Wavelet Transform decomposition level five and six
Perbandingan Ekstraksi Ciri Haar Wavelet dan Log-Gabor Wavelet pada Pengenalan Iris Mata dengan VFI5
No full textA biometric system provides automatic identification based on unique feature or characteristic by the individual. As an example, the iris pattern of human eye is used in biometric system. Iris pattern is more stable and unique so it is suitable for biometric system. This research aims to compare dimension of Haar wavelet and Log-Gabor wavelet as feature extraction on iris recognition using VFI5 algorithm and provides implementation for recognizing eye images based on VFI5 algorithm. The eye images in this research are taken from CASIA dataset. The system uses Canny edge detector for image segmentation and Daugman’s Rubber Sheet Model to normalize the result to constant interval. Feature extraction uses Haar wavelet with 3 different levels and iris recognition with VFI5 algorithm with 3-cross fold validation. As a result, the best recognition of testing data is obtained from combining the votes from the left and right eyes. Less dimensions and features are obtained when using Haar wavelet than Log-Gabor wavelet
Analisis Grafologi Berdasarkan Huruf a dan t Menggunakan Algoritme Voting Feature Intervals 5
No full textGraphology is a scientific method to identify, evaluate and understand human personality through the strokes and patterns revealed by handwriting. Handwriting will indicate the true personality including emotional, fear, honesty, defenses and many others. People who studied graphology called grapohologist. Graphologist has a subjective assessment in handwriting analyzing. Different graphologist can analyze the same handwriting but the result will shown in a different way. The accuracy of handwriting depends on the graphologist ability. Therefore, the computer technology will be needed to apply the science of graphology to help graphologist in analyzing handwriting. This research is concern to developed a model of handwriting analysis based on letter a and letter t using Voting Feature Intervals 5 algorithm. Image as an input for VFI5 algorithm and personality as an output. The result showed that letter a easier to be recognized than letter t
- …
