Hybrid Fuzzy Clustering dan Bagging Support Vector Machine untuk Klasifikasi Tingkat Depresi dengan Terapi Mindfulness Berdasarkan Electroencephalogram

Karimah, Saffana (2025) Hybrid Fuzzy Clustering dan Bagging Support Vector Machine untuk Klasifikasi Tingkat Depresi dengan Terapi Mindfulness Berdasarkan Electroencephalogram. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6003231030-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Depresi merupakan penyakit mental yang ditandai dengan suasana hati yang buruk secara terus menerus, kesedihan, putus asa, gangguan kognitif, dan gangguan tidur. Kondisi ini menjadi perhatian serius karena memiliki dampak signifikan terhadap kualitas hidup seseorang. Salah satu psikoterapi untuk penanganan kondisi depresi ialah Mindfulness-Based Cognitive Therapy (MBCT) yaitu praktik kesadaran penuh yang membantu individu mengarahkan fokus pada pemahaman terhadap kondisi yang dialaminya sehingga memunculkan penerimaan diri. Evaluasi efektivitas MBCT pada pasien depresi dapat dilakukan dengan menganalisis perbedaan sinyal gelombang otak (Electroencephalogram atau EEG) pada responden yang sama sebelum dan setelah terapi. EEG memiliki bentuk yang kompleks sehingga perlu diproses lebih lanjut untuk memperoleh informasi yang mendalam. Pada penelitian ini, sinyal EEG diproses menggunakan filtrasi Finite Impulse Response, dilanjutkan ekstraksi fitur. Fitur hasil ekstraksi pada data pre-treatment digunakan untuk mengelompokkan responden depresi menggunakan Revised Fuzzy C-Means Clustering. Proses clustering menghasilkan dua kelompok depresi pada tiga skenario channel EEG yaitu FP1, FP2, dan kombinasi FP1 dan FP2. Evaluasi clustering menunjukkan channel FP2 memiliki kualitas clustering terbaik, menandakan sinyal EEG pada channel FP2 memiliki karakteristik paling diskriminatif dalam membedakan kondisi depresi. Hasil clustering kemudian digabungkan dengan data responden normal dan digunakan untuk membangun model klasifikasi menggunakan Balanced Bagging Support Vector Machine (SVM). Model terbaik diperoleh pada channel FP1 dengan F1-Score Macro 0,8358. Model ini digunakan untuk memprediksi kondisi mental responden saat perekaman kedua, baik pada responden perlakuan (responden yang mengikuti terapi MBCT) maupun pada responden kontrol (responden yang tidak mengikuti terapi MBCT). Berdasarkan analisis, terapi MBCT secara umum terbukti memberikan dampak positif terhadap perbaikan kondisi depresi responden, di mana 55% responden kelompok perlakuan mengalami perbaikan. Persentase ini lebih besar dibandingkan dengan kelompok kontrol yang hanya 25% responden yang mengalami perbaikan kondisi.
==================================================================================================================================
Depression is a mental illness characterized by a persistently low mood, sadness, hopelessness, cognitive disturbances, and sleep disorders. This condition is a serious concern due to its significant impact on an individual’s quality of life. One of the psychotherapeutic approaches for managing depression is Mindfulness-Based Cognitive Therapy (MBCT), which involves the practice of full awareness to help individuals focus on understanding their mental state, thus fostering self-acceptance. The effectiveness of MBCT in patients with depression can be evaluated by analyzing the differences in brainwave signals (Electroencephalogram or EEG) recorded from the same respondents before and after therapy. Due to the complex nature of EEG signals, further processing is required to extract meaningful information. In this study, EEG signals were processed using Finite Impulse Response (FIR) filtering, followed by feature extraction. The extracted features from pre-treatment data were used to cluster depressed respondents using the Revised Fuzzy C-Means (RFCM) Clustering method. The clustering process generated two depression groups across three EEG channel scenarios: FP1, FP2, and a combination of FP1 and FP2. Clustering evaluation indicated that the FP2 channel produced the best clustering quality, suggesting that EEG signals from this channel had the most discriminative characteristics for distinguishing depressive conditions. The clustering results were then combined with normal respondent data and used to develop a classification model using the Balanced Bagging Support Vector Machine (SVM). The best model was obtained from the FP1 channel with a Macro F1-Score of 0.8358. This model was then used to predict the mental state of respondents during the second EEG recording, both for the treatment group (those who received MBCT) and the control group (those who did not). Based on the analysis, MBCT was generally proven to have a positive effect on improving depressive symptoms, with 55% of the treatment group showing improvement— a higher percentage compared to the control group, in which only 25% of respondents experienced improvement.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Depresi, Electroencephalogram, Ensemble SVM, Fuzzy Clustering, Mindfulness
Subjects:	H Social Sciences > HA Statistics H Social Sciences > HA Statistics > HA29 Theory and method of social science statistics R Medicine > RM Therapeutics. Pharmacology
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis
Depositing User:	Saffanah Karimah
Date Deposited:	06 Aug 2025 03:26
Last Modified:	06 Aug 2025 03:26
URI:	http://repository.its.ac.id/id/eprint/127722

Actions (login required)

View Item