Electronic Theses and Dissertation

Abstract- Autism Spectrum Disorder (ASD) atau sindrom neurodevelopmental merupakan gangguan pada perkembangan saraf otak yang menyebabkan penderita mengalami penurunan interaksi sosial, kemampuan berkomunikasi, serta pengekspresian emosi. Sindrom autis dapat dideteksi dengan menggunakan Elektroensefalogram (EEG). Penelitian ini memanfaatkan EEG penyandang autis guna mendukung studi klasifikasi skema machine learning dengan menghasilkan akurasi terbaik. Metode Linear Discriminant Analysis (LDA) merupakan salah satu teknik machine learning yang digunakan untuk mengklasifikasikan sinyal EEG autis dan normal. LDA dipilih karena dapat memaksimalkan jarak antar kelas dan meminimalkan jumlah scatter dengan memanfaatkan fungsi between dan within-class. Metode ini akan dikombinasikan dengan metode lainnya yaitu Independent Component Analysis (ICA) dan Discrete Wavelet Transform (DWT) untuk meningkatkan akurasi sistem. Banyak sampel yang digunakan berjumlah 16 data EEG yang merekam 8 anak autis dan 8 anak normal. Sinyal dari dataset dilakukan filterisasi artifak menggunakan Independent Component Analysis (ICA) dan didekomposisi sebanyak 3 level melalui Discrete Wavelet Transform (DWT) kemudian diklasifikasi menggunakan teknik Linear Discriminant Analysis (LDA). Berdasarkan hasil penelitian, akurasi terbaik yang diperoleh sebesar 99% dengan bantuan Confusion Matrix.


Kata kunci: Autis, Elektroensefalogram, Linear Discriminant Analysis, Independent Component Analysis, Discrete Wavelet Transform

Abstract- Autism Spectrum Disorder (ASD) is a neurodevelopment syndrome that causes sufferers to experience a decrease in social interaction, communication skills, and emotional expression. Autism syndrome can be detected by using an electroencephalogram (EEG). This study utilizes EEG of autism people to support the classification study of machine learning schemes to produce the best accuracy. The Linear Discriminant Analysis (LDA) method is a machine learning technique used to classify autism and normal EEG signals. LDA was chosen because it can maximize the distance between classes and minimize the number of scatters by utilizing between and within-class functions. This method will be combined with other methods, namely Independent Components Analysis (ICA) and Discrete Wavelet Transform (DWT) to improve the accuracy system. Dataset EEG collated from 16 children that recorded 8 autism and 8 normal. The signals from the dataset were filtered by artifacts using Independent Component Analysis (ICA) and decomposed by 3 levels through Discrete Wavelet Transform (DWT) then classified using the Linear Discriminant Analysis (LDA) technique. Based on the results, the best accuracy obtained is 99% with the help of the Confusion Matrix. Keywords: Autism, Electroencephalogram, Linear Discriminant Analysis, Independent Component Analysis, Discrete Wavelet Transform