Stochastic approximation for learning rate optimization for generalized relevance learning vector quantization

Author(s):  
Daniel W. Steeneck ◽  
Trevor J. Bihl
2020 ◽  
Vol 2020 ◽  
pp. 1-12
Author(s):  
Trevor J. Bihl ◽  
Todd J. Paciencia ◽  
Kenneth W. Bauer ◽  
Michael A. Temple

Radio frequency (RF) fingerprinting extracts fingerprint features from RF signals to protect against masquerade attacks by enabling reliable authentication of communication devices at the “serial number” level. Facilitating the reliable authentication of communication devices are machine learning (ML) algorithms which find meaningful statistical differences between measured data. The Generalized Relevance Learning Vector Quantization-Improved (GRLVQI) classifier is one ML algorithm which has shown efficacy for RF fingerprinting device discrimination. GRLVQI extends the Learning Vector Quantization (LVQ) family of “winner take all” classifiers that develop prototype vectors (PVs) which represent data. In LVQ algorithms, distances are computed between exemplars and PVs, and PVs are iteratively moved to accurately represent the data. GRLVQI extends LVQ with a sigmoidal cost function, relevance learning, and PV update logic improvements. However, both LVQ and GRLVQI are limited due to a reliance on squared Euclidean distance measures and a seemingly complex algorithm structure if changes are made to the underlying distance measure. Herein, the authors (1) develop GRLVQI-D (distance), an extension of GRLVQI to consider alternative distance measures and (2) present the Cosine GRLVQI classifier using this framework. To evaluate this framework, the authors consider experimentally collected Z-wave RF signals and develop RF fingerprints to identify devices. Z-wave devices are low-cost, low-power communication technologies seen increasingly in critical infrastructure. Both classification and verification, claimed identity, and performance comparisons are made with the new Cosine GRLVQI algorithm. The results show more robust performance when using the Cosine GRLVQI algorithm when compared with four algorithms in the literature. Additionally, the methodology used to create Cosine GRLVQI is generalizable to alternative measures.


Author(s):  
Eko Arianto ◽  
Laifa Rahmawati

One of the lessons for mental disorder students in Special Schools is practicum lessons in the form of vocational education. This lesson uses equipment that requires prudence. Mental disorder students have characteristics that are low memory and move based on intuition. Teachers should pay extra attention especially to detect student behavior during the learning. This detection is needed for learning to take place smoothly and students are safe from the dangers around the practicum place. Teacher's feedback on the detection obtained in the form of a warning from the teacher. This study is expected to be useful for providing a special detection pattern for students to assist teachers by providing feedback in the form of warnings using natural motion detection technology. This research was conducted using Kinect as data input and data was processed using artificial neural network and Learning Vector Quantization method. The dangerous attitude used in the test is the attitude of standing at the time of drilling position. The data used by training is 126 data and do training using LVQ. At the LVQ training stage, the training was conducted with parameter of Learning Rate 0,05, maximum Iteration 44, reduction of learning rate 0.01, and Learning rate minimum 0,02.


2020 ◽  
Vol 6 (1) ◽  
pp. 28-35
Author(s):  
Ery Murniyasih ◽  
Luluk Suryani

Penelitian ini bertujuan : (1). Membuat suatu aplikasi untuk identifikasi jenis penyakit pada tanaman padi berdasarkan bentuk bercak daun padi.;(2). Menerapkan metode Learning Vector Quantization (LVQ) pada identifikasi penyakit tanaman padi. Pada tahapan learning dan testing pada LVQ citra diproses menjadi Grayscale, Thresholding, dan segmentasi. Di tahap pelatihan, metode LVQ digunakan untuk menentukan bobot, target error, max epoch, dan laju pelatihan (Learning rate). Data yang dijadikan sebagai input adalah citra identifikasi jenis penyakit tanaman padi berdasarkan bentuk bercak daun padi  yaitu dengan ukuran piksel 95x35 dan berekstensi BITMAP (.bmp). Standar keberhasilan sistem identifikasi ini adalah menghitung nilai Termination Error Rate dan tingkat keakuratan dalam identifikasi bentuk bercak daun. Dari simulasi ini diperoleh struktur Jaringan Syaraf Tiruan dengan jumlah nilai learning rate 0,02 dan jumlah epoch sebesar 5 kali. Sistem yang terbentuk mampu mengenali citra yang berisi bentuk bercak daun yang digunakan sebagai bobot dengan nilai keakuratan optimum yaitu 73,33% dengan komposisi penyakit bercak coklat (BC) 20 %, Blast  20 % dan cercak cercospora 33,33%.


2018 ◽  
Vol 15 (2) ◽  
pp. 144
Author(s):  
Elvia Budianita Budianita

Trimester I adalah masa dimana 3 bulan pertama kehamilan yakni 0 sampai 12 minggu awal kehamilan. Pada masa ini tubuh ibu akan banyak mengalami perubahan seiring berkembangnya janin. Pada ibu-ibu hamil pada fase trimester I terkadang ditemukan beberapa gangguan kehamilan yaitu, Abortus, Anemia Kehamilan, Hiperemesis Gravidarum tingkat I, Hiperemesis Gravidarum tingkat II, Kehamilan Ektopik, dan Mola hidatidosa. Untuk membantu pasien dalam mengenali gangguan kehamilan pada trimester I ini maka peneliti berinisiatif merancang suatu sistem yang menerapkan konsep jaringan syaraf tiruan dengan metode LVQ 2 (Learning Vector Quantization) dalam mengenali gangguan kehamilan trimester I berdasarkan gejala gangguan kehamilan trimester I. Ada 41 gejala penyakit, dan 6 penyakit sebagai data masukan. Sistem akan mengklasifikasikan penyakit dengan proses pembelajaran dan pengujian ke dalam 6 jenis penyakit, berdasarkan pengujian metode LVQ2 cukup baik di terapkan dalam pengenalan pola gejala gangguan kehamilan, di buktikan dari hasil pengujian yang di lakukan menggunakan window 0.1, 0.3, 0.5, dan 0, data latih 90 dan data uji 18 didapat akurasi terbaik 100% dan rata-rata akurasi 97.68%  dengan nilai parameter pembelajaran algoritma learning rate = 0.02, 0.04, 0.06, pengurangan learning rate = 0.1, minimal learning rate = 0.01 dan nilai window (ε) =0.1, 0.3, 0.5, dan 0. Nilai w juga mempengaruhi akurasi. Kata Kunci:  Gangguan Kehamilan Trimester I, Learning Vector Quantization 2, Window


2019 ◽  
Vol 5 (2) ◽  
pp. 123
Author(s):  
Erwin Yudi Hidayat ◽  
Muhammad Farhan Radiffananda

Tanda tangan merupakan salah satu biometrik pada karakteristik perilaku yang digunakan untuk mengenali seseorang sebagai sistem identifikasi. Meskipun unik, banyak terjadi kasus tanda tangan yang disalahgunakan dengan cara dipalsukan. Tidak mudah mengenali tanda tangan yang palsu dengan tanda tangan asli. Penelitian ini menerapkan algoritma Learning Vector Quantization, deteksi tepi Sobel, dan ekstraksi fitur Local Binary Pattern untuk mengidentifikasi tanda tangan. Hasil penelitian menunjukkan, jumlah data citra, iterasi, dan learning rate mempengaruhi akurasi dan waktu proses identifikasi. Dari percobaan yang dilakukan pada parameter yang berbeda-beda, akurasi yang didapat adalah 68% pada data latih dan pada data uji sebesar 54,6%.Kata kunci—identifikasi, Learning Vector Quantization, tanda tangan, pengenalan pola


2020 ◽  
Vol 4 (2) ◽  
pp. 75-85
Author(s):  
Chrisani Waas ◽  
D. L. Rahakbauw ◽  
Yopi Andry Lesnussa

Artificial Neural Network (ANN) is an information processing system that has certain performance characteristics that are artificial representatives based on human neural networks. ANN method has been widely applied to help human performance, one of which is health. In this research, ANN will be used to diagnose cataracts, especially Congenital Cataracts, Juvenile Cataracts, Senile Cataracts and Traumatic Cataracts based on the symptoms of the disease. The ANN method used is the Learning Vector Quantization (LVQ) method. The data used in this research were 146 data taken from the medical record data of RSUD Dr. M. Haulussy, Ambon. The data consists of 109 data as training data and 37 data as testing data. By using learning rate (α) = 0.1, decrease in learning rate (dec α) = 0.0001 and maximum epoch (max epoch) = 5, the accuracy rate obtained is 100%.


2015 ◽  
Vol 2 (2) ◽  
pp. 128
Author(s):  
Fajar Rohman Hariri ◽  
Ema Utami ◽  
Armadyah Amborowati

Data berukuran besar yang sudah disimpan jarang digunakan secara optimal karena manusia seringkali tidak memiliki waktu dan kemampuan yang cukup untuk mengelolanya. Data bervolume besar seperti data teks, jauh melampaui kapasitas pengolahan manusia yang sangat terbatas. Kasus yang disoroti adalah data abstrak tugas akhir mahasiswa jurusan teknik informatika Universitas Trunojoyo Madura. Dokumen tugas akhir oleh mahasiswa terkait hanya diupload pada SIMTAK (Sistem Informasi Tugas Akhir) dan pelabelan bidang minat penelitian dilakukan manual oleh mahasiswa tersebut, sehingga akan ada kemungkian saat mahasiswa mengisi bidang minat tidak sesuai. Untuk menanggulangi hal tersebut, diperlukan adanya mekanisme pelabelan dokumen secara otomatis, untuk meminimalisir kesalahan. Pada penelitian kali ini dilakukan klasifikasi dokumen abstrak tugas akhir menggunakan metode Learning Vector Quantization (LVQ). Data abstrak diklasifikasikan menjadi 3 yaitu SI RPL (Sistem Informasi – Rekayasa Perangkat Lunak), CAI (Computation – Artificial Intelligence) dan Multimedia. Dari berbagai ujicoba yang dilakukan didapatkan hasil metode LVQ berhasil mengenali 90% data abstrak, dengan berhasil mengenali 100% bidang minat SI RPL dan CAI, dan hanya 70% untuk bidang minat Multimedia. Dengan kondisi terbaik didapatkan dengan parameter reduksi dimensi 20% dan nilai learning rate antara 0,1-0,5.Huge size of data that have been saved are rarely used optimally because people often do not have enough time and ability to manage. Large volumes of data such as text data, exceed human processing capacity. The case highlighted was the final project abstract data from informatics engineering student Trunojoyo University. Documents abstract just uploaded on SIMTAK (Final Project Information System) and the labeling of the areas of interest of research is done manually by the student, so that there will be a possibility to fill the field of interest while the student is not appropriate. To overcome this, we need a mechanism for labeling a document automatically, to minimize errors. In the present study conducted abstract document classification using Learning Vector Quantization (LVQ). Abstract data classified into three class, SI RPL, CAI and Multimedia. Of the various tests carried out showed that LVQ method successfully recognize 90% of abstract data, to successfully identify 100% interest in the field of RPL SI and CAI, and only 70% for areas of interest Multimedia. With the best conditions obtained with the parameter dimension reduction of 20% and the value of learning rate between 0.1-0.5.


Sign in / Sign up

Export Citation Format

Share Document