11 January 2012

Machine Learning - complete course notes

Stanford's machine learning course presented by Professor Andrew Ng and originally posted on the ml-class.org website during the fall 2011 semester
Machine Learning - complete course notes:

'via Blog this'

Application_Example_OCR

Stanford
18_Application_Example_OCR:

'via Blog this'

Clustering

Stanford
13_Clustering:

'via Blog this'

Machine_Learning_System_Design

Stanford
11_Machine_Learning_System_Design:

'via Blog this'

Advice_for_applying_machine_learning

Standford
10_Advice_for_applying_machine_learning:

'via Blog this'

Neural_Networks_Learning

Stanford
09_Neural_Networks_Learning:

'via Blog this'

Neural_Networks_Representation

Stanford
08_Neural_Networks_Representation:

'via Blog this'

Logistic_Regression -- Stanford

06_Logistic_Regression:

'via Blog this'

04 January 2012

Fuzzy Hidden Markov Models For Indonesian Speech Classification

World Congress of International Fuzzy System Association and Asia Fuzzy System Society International Conference, Surabaya-Bali, 2011
Fuzzy Hidden Markov Models For Indonesian Speech Classification
Intan Nurma Yulita 1), The Houw Liong 2), Adiwijaya 3)
Graduate Faculty, Telkom Institute of Technology
Jalan Telekomunikasi No.1, Dayeuhkolot, Jawa Barat, Indonesia
Email: intanurma@gmail.com, houwthee@yahoo.co.id, adw@ittelkom.ac.id
Abstract
Indonesia has a lot of tribe, so that there are a lot of dialects. Speech classification is difficult if the database uses speech signals from various people who have different characteristics because of gender and dialect. The different characteristics will influence frequency, intonation, amplitude, and period of the speech. It makes the system be trained for the various templates reference of speech signal. Therefore, this study has been developed for Indonesian speech classification. This study designs the solution of the different characteristics for Indonesian speech classification. The solution combines Fuzzy on Hidden Markov Models. The new design of fuzzy Hidden Markov Models will be proposed in this study. The models will consist of Fuzzy C-Means Clustering which will be designed to substitute the vector quantization process and a new forward and backward method to handle the membership degree of data. The result shows FHMM is better than HMM.
Keywords: Fuzzy, Hidden Markov Models, Indonesian, Speech, Classification, Clustering

THE PREDICTION OF DENGUE HAEMORRAGIC FEVER (DHF) IN CIMAHI USING HYBRID GENETIC ALGORITHM AND FUZZY LOGIC

International Conference on Informatics and Computational Intelligence, Bandung, 2011
THE PREDICTION OF DENGUE HAEMORRAGIC FEVER (DHF) IN CIMAHI USING HYBRID GENETIC ALGORITHM AND FUZZY LOGIC
Fhira Nhita,ST 1), Prof.Thee Houw Liong 2), Shaufiah,MT 3)
1,3 Informatics Program Study in Telkom Institute of Technology 2Bandung Institute of Technology
1)fhiranhita@yahoo.com, 2)thehl007@gmail.com, 3)shaufiah@gmail.com
Abstract
The incidence of Dengue Haemorrhagic Fever (DHF) is a national health problem in Indonesia. Every year dengue morbidity is still high. Particularly in Cimahi, one of the city in West Java province where the morbidity rate (Incidence Rate) 2005 to 2010 in the above national standard. Many factors can affect the incidence of dengue, among others, climate and living behavioral. Therefore, the development of DHF Prediction System which is associated with a climate that is expected to help provide information for the Department of Health about dengue risk prediction in the coming year, so the Health Ministry can take preventive action to reduce morbidity of DHF. The Prediction System that was built with a hybrid algorithm which Genetic Algorithms and Fuzzy Logic is able to obtain 100% testing accuracy in predicting the condition of dengue in the first 6 months in 2009 and 2010 in North Cimahi and Central Cimahi. While in South Cimahi, the prediction results obtained in the first 6 months of 2009 amounted to 100% but in 2010 a decline in accuracy.
Keywords: dengue haemorrhagic fever, cimahi, genetic algorithm, fuzzy logic, prediction system
Abstrak
Kejadian Demam Berdarah Dengue (DBD) merupakan masalah nasional di bidang kesehatan. Setiap tahun angka kesakitan DBD masih tinggi. Khususnya di Cimahi, salah satu kota di provinsi Jawa Barat dimana angka kesakitan (Incidence Rate) tahun 2005 hingga 2010 di atas standar yang ditentukan Departemen Kesehatan RI. Banyak faktor yang mempengaruhi kejadian DBD, antara lain iklim dan perilaku hidup bersih dan sehat (PHBS). Oleh karena itu, dalam penelitian ini, dibangun Sistem Prediksi demam berdarah yang dikaitkan dengan iklim, yang diharapkan bisa membantu memberikan informasi bagi Departemen Kesehatan tentang prediksi resiko DBD di tahun yang akan datang, sehingga Departemen Kesehatan dapat mengambil langkah preventif untuk mengurangi angka kesakitan DBD. Sistem Prediksi yang dibangun dengan hybrid algorithm yaitu Algoritma Genetika dan Fuzzy Logic mampu menghasilkan akurasi testing 100% dalam memprediksi kondisi DBD di 6 bulan pertama pada tahun 2009 dan 2010 di kecamatan Cimahi Utara dan Cimahi Tengah. Sedangkan pada Cimahi Selatan diperoleh hasil prediksi 6 bulan pertama di tahun 2009 sebesar 100% tetapi pada tahun 2010 terjadi penurunan akurasi. Kata kunci: demam berdarah, cimahi, algoritma genetika, logika fuzzy, system prediksi

Data Mining , IT Telkom

Intan Nurma Yulita:

'via Blog this'