Agus Harjoko
Departemen Ilmu Komputer dan Elektronika, FMIPA UGM, Yogyakarta

Published : 11 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 4 Documents
Search
Journal : IJCCS (Indonesian Journal of Computing and Cybernetics Systems)

Pengenalan Ucapan Suku Kata Bahasa Lisan Menggunakan Ciri LPC, MFCC, dan JST Abriyono Abriyono; Agus Harjoko
IJCCS (Indonesian Journal of Computing and Cybernetics Systems) Vol 6, No 2 (2012): July
Publisher : IndoCEISS in colaboration with Universitas Gadjah Mada, Indonesia.

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.22146/ijccs.2149

Abstract

AbstrakSuara adalah salah satu alat komunikasi antar manusia yang efektif dan digemari. Selain sebagai alat komunikasi antar manusia, suara manusia telah digunakan sebagai alat komunikasi antara manusia dan komputer (mesin). Penelitian menggunakan suara sebagai alat komunikasi manusia dan mesin telah banyak dilakukan dengan menggunakan berbagai bahasa. Bahkan ada beberapa penelitian yang telah menghasilkan kemampuan pengenalan yang baik dan dikomersilkan (menggunakan bahasa Inggris). Bagaimana dengan penelitian pengenalan suara menggunakan Bahasa Indonesia? Peneliti mengamati penelitian pengenalan ucapan kata dalam Bahasa Indonesia masih minim dan cakupan jumlah katanya pun masih kecil. Oleh karena itu, pada penelitian ini, peneliti melakukan pengenalan ucapan kata Bahasa Indonesia. Pengenalan ucapan kata Bahasa Indonesia dilakukan dengan memecah kata Bahasa Indonesia ke dalam bentuk suku kata bahasa lisan. Pemecahan ke dalam bentuk lafal kata diharapkan mampu mengurangi jumlah kata yang sangat besar, namun tetap mengakomodasi seluruh kata yang dalam Bahasa Indonesia. Total jumlah lafal kata yang ditemukan oleh peneliti adalah 1741 suku kata bahasa lisan. Peneliti membagi sistem dalam 4 bagian besar, yakni proses perekaman, pre-processing, ekstraksi ciri, dan pengenalan. Pada proses perekaman digunakan frekuensi 11025 Hz, Mono, 8 bit. Pada pre-processing digunakan proses bantuan seperti pre-emphasis, segmentasi, framing, dan windowing. Sedangkan untuk ekstraksi ciri dan pengenalan digunakan ciri LPC/MFCC dan identifier jaringan syaraf tiruan backpropagation. Hasil pengenalan dengan pendekatan yang dibangun menunjukkan hasil yang belum memuaskan, yakni dengan kemampuan pengenalan terbaik sebesar 0.65% dengan ciri MFCC. Kata kunci—pengenalan kata Bahasa Indonesia, LPC, MFCC, JST, backpropagation. Abstract Voice is one of effective and convinienced communication’s medium among human. However, the used of voice is not only for communication among human but also has another role nowadays. Voice becomes communication medium for human and computer (machine). One of its application is speech to text application. Some of speech to text research already claimed good accuracy for some languages. How about Indonesian language? The research for Indonesian word recognition was still at low amount. The word used for research was at small amount too. Because of some of the reasons, researcher focus on Indonesian word recognition in this research. This research will divide the word into the speech syllable. The aim for the dividing system is to reduce the large amount of the word, but still cover all of the word. We found and used 1741 speech syllables. For managing the recognition, we used several approaches. The approaches are 11025 Hz, Mono, 8 bit for recording, pre-emphasized, segmentation, framing, and windowing for pre-processing, LPC and MFCC for the features, and back-propagation neural network for the identifier. The result using this approach was not reached good performance. The best result performed 0.65% by using MFCC feature. Keywords—Indonesian’s syllable recognition, LPC, MFCC, neural network, backpropagation
Sistem Informasi Geografis Risiko Kemunculan Rip Current Menggunakan Decision Tree C4.5 Made Leo Radhitya; Agus Harjoko
IJCCS (Indonesian Journal of Computing and Cybernetics Systems) Vol 10, No 2 (2016): July
Publisher : IndoCEISS in colaboration with Universitas Gadjah Mada, Indonesia.

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.22146/ijccs.15949

Abstract

One of the dangers that occur at the beach is rip current. Rip current poses significant danger for beachgoers. This paper proposes a method to predict the rip current's occurence risk by using decision tree generated using C4.5 algorithm. The output from the decision tree is rip current's occurrence risk. The case study for this research is the beach located at Rote Island, Rote Ndao, Nusa Tenggara Timur. Evaluation result shows that the accuracy is 0.84, and the precision is 0.61. The average recall value is 0.68 and the average F-measure is 0.59 in the range 0 to 1.
Klasifikasi Varietas Cabai Berdasarkan Morfologi Daun Menggunakan Backpropagation Neural Network Kharis Syaban; Agus Harjoko
IJCCS (Indonesian Journal of Computing and Cybernetics Systems) Vol 10, No 2 (2016): July
Publisher : IndoCEISS in colaboration with Universitas Gadjah Mada, Indonesia.

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.22146/ijccs.16628

Abstract

Compared with other methods of classifiers such as cellular and molecular biological methods, using the image of the leaves become the first choice in the classification of plants. The leaves can be characterized by shape, color, and texture; The leaves can have a color that varies depending on the season and geographical location. In addition, the same plant species also can have different leaf shapes. In this study, the morphological features of leaves used to identify varieties of pepper plants. The method used to perform feature extraction is a moment invariant and basic geometric features. For the process of recognition based on the features that have been extracted, used neural network methods with backpropagation learning algorithm. From the neural-network training, the best accuracy in classifying varieties of chili with minimum error 0.001 by providing learning rate 0.1, momentum of 0.7, and 15 neurons in the hidden layer foreach of various feature. To conduct cross-validation testing with k-fold tehcnique, obtained classification accuracy to be range of 80.75%±0.09% with k=4.
Case-Based Reasoning for Stroke Disease Diagnosis Nelson Rumui; Agus Harjoko; Aina Musdholifah
IJCCS (Indonesian Journal of Computing and Cybernetics Systems) Vol 12, No 1 (2018): January
Publisher : IndoCEISS in colaboration with Universitas Gadjah Mada, Indonesia.

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.22146/ijccs.26331

Abstract

Stroke is a type of cerebrovascular disease that occurs because blood flow to the brain is disrupted. Examination of stroke accurately using CT scan, but the tool is not always available, so it can be done by the Siriraj Score. Each type of stroke has similar symptoms so doctors should re-examine similar cases prior to diagnosis. The hypothesis of the Case-based reasoning (CBR) method is a similar problems having similar solution.This research implements CBR concept using Siriraj score, dense index and Jaccard Coeficient method to perform similarity calculation between cases.The test is using k-fold cross validation with 4 fold and set values of threshold (0.65), (0.7), (0.75), (0.8), (0.85), (0.9), and (0.95). Using 45 cases of data test  and 135 cases of case base. The test showed that threshold of 0.7 is suitable to be applied in sensitivity (89.88%) and accuracy (84.44% for CBR using indexing and 87.78% for CBR without indexing). Threshold of 0.65 resulted high sensitivity  and accuracy but showed many cases of irrelevant retrieval results. Threshold (0.75), (0.8), (0.85), (0.9) and (0.95) resulted in sensitivity (65.48%, 59.52%, 5.95%, 3,57% and 0%) and accuracy of CBR using indexing (61.67%, 55.56%, 5.56%, 3.33%, and 0%) and accuracy of CBR without indexing (62.78% 56.67%, 55.56%, 5.56%, 3.33%, and 0%).