Pratasik, Stralen (2026) Studi Neuromarketing Berbasis Sinyal EEG Menggunakan Video Reklame Produk. Doctoral thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 7022222003-Doctoral.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
7022222003-Doctoral.pdf - Accepted Version Restricted to Repository staff only Download (7MB) | Request a copy |
Abstract
Neuromarketing merupakan bidang interdisipliner yang menggabungkan ilmu neurosains dan perilaku konsumen untuk memahami bagaimana aktivitas otak memengaruhi proses pengambilan keputusan pembelian. Aktivitas tersebut dapat diukur secara objektif melalui pola sinyal Electroencephalography yang merepresentasikan dinamika gelombang otak ketika individu menerima stimulus visual. Penelitian ini bertujuan untuk menyelidiki respons otak konsumen terhadap tayangan video reklame produk menggunakan analisis sinyal EEG guna memahami keterlibatan emosi dan atensi dalam proses pembentukan minat pembelian. Eksperimen dilakukan terhadap 28 partisipan dengan menggunakan dua video reklame sepatu olahraga lokal Indonesia. Perekaman EEG dilakukan pada enam kanal utama, yaitu Fp1, Fp2, F7, F8, O1, dan O2, yang merepresentasikan aktivitas frontal (emosi) dan oksipital (atensi). Data EEG yang diperoleh diproses melalui tahap pra-pemrosesan kemudian dilakukan dekomposisi pita frekuensi menggunakan dua pendekatan, yaitu dengan Discrete Wavelet Transform berbasis Daubechies (db4) dan dekomposisi pita frekuensi menggunakan Butterworth filter. Kemudian dilakukan ekstraksi fitur domain waktu, domain frekuensi, dan domain waktu–frekuensi. Selanjutnya, seleksi fitur dilakukan menggunakan tiga metode utama, yaitu Pearson Correlation, Linear Discriminant Analysis, dan Mutual Information, sedangkan klasifikasi dilakukan menggunakan lima algoritma kecerdasan buatan: Support Vector Machine, K-Nearest Neighbor, Random Forest, Naïve Bayes, dan Decision Tree. Hasil penelitian menunjukkan bahwa pita gamma dan beta memiliki kontribusi paling signifikan dalam membedakan kondisi high buying intention dan low buying intention. Kanal O2 menunjukkan keterlibatan dominan dalam pemrosesan visual dan perhatian, sedangkan kanal Fp2 dan F8 berperan dalam aktivitas emosional dan evaluatif terhadap video reklame. Dari sisi performa model, Random Forest menghasilkan akurasi tertinggi, mencapai 98.0 % ± 0.3 % pada analisis pita gamma dan 93% pada analisis domain waktu–frekuensi. Sementara itu, Mutual Information terbukti sebagai metode seleksi fitur paling efisien, hanya membutuhkan 20 fitur untuk mencapai akurasi di atas 90% dengan tingkat redundansi rendah (14,2%). Secara keseluruhan, penelitian ini berhasil membangun model prediksi intensi pembelian berbasis EEG yang efisien, akurat, dan valid. Kombinasi analisis domain waktu–frekuensi EEG, seleksi fitur adaptif, dan model kecerdasan buatan dapat digunakan sebagai pendekatan objektif dan kuantitatif dalam mengevaluasi efektivitas video reklame berdasarkan aktivitas otak konsumen. Pendekatan ini berpotensi menjadi dasar pengembangan sistem neuromarketing berbasis EEG untuk mendukung strategi pemasaran berbasis data neurokognitif di masa depan.
=====================================================================================================================================
Neuromarketing is an interdisciplinary field that integrates neuroscience and consumer behavior to understand how brain activity influences the decision-making process in purchasing behavior. This activity can be objectively measured through the patterns of Electroencephalography signals, which represent the dynamics of brain waves when individuals are exposed to visual stimuli such as product advertising videos. This study aims to investigate consumers’ brain responses to product advertising videos using EEG signal analysis to understand the involvement of emotion and attention in the formation of buying intention. The experiment was conducted with 28 participants using two local Indonesian sports shoe commercials. EEG recording was performed using six main channels, Fp1, Fp2, F7, F8, O1, and O2 representing frontal (emotional) and occipital (attentional) brain activities. The acquired EEG data underwent preprocessing, followed by frequency band decomposition using two approaches: Discrete Wavelet Transform based on Daubechies (db4) and frequency band segmentation using Butterworth filtering. Feature extraction was then carried out across the time, frequency, and time–frequency domains. Subsequently, feature selection was performed using three main methods: Pearson Correlation, Linear Discriminant Analysis, and Mutual Information, while classification was conducted using five artificial intelligence algorithms: Support Vector Machine, K-Nearest Neighbor, Random Forest, Naïve Bayes, and Decision Tree.
The results revealed that the gamma and beta bands made the most significant contributions in distinguishing between high buying intention and low buying intention conditions. The O2 channel demonstrated dominant involvement in visual and attentional processing, while the Fp2 and F8 channels were associated with emotional and evaluative activities toward the advertisements. In terms of model performance, Random Forest achieved the highest accuracy around 98.0 % ± 0.3 % in the gamma band analysis and 93% in the time–frequency domain analysis. Meanwhile, Mutual Information proved to be the most efficient feature selection method, requiring only 20 features to achieve accuracy above 90% with a low redundancy rate (14.2%). Overall, this study successfully developed an EEG-based buying intention prediction model that is efficient, accurate, and valid. The combination of EEG time–frequency domain analysis, adaptive feature selection, and artificial intelligence models can serve as an objective and quantitative approach to evaluating advertisement effectiveness based on consumers’ brain activity. This approach holds potential as a foundation for developing EEG-based neuromarketing systems to support data-driven neurocognitive marketing strategies in the future.
| Item Type: | Thesis (Doctoral) |
|---|---|
| Uncontrolled Keywords: | atensi, electroencephalography, emosi, video reklame, attention, electroencephalography, emotions, memory, neuromarketing |
| Subjects: | Q Science > QP Physiology > Q376.5 Electroencephalography (EEG) |
| Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20001-(S3) PhD Thesis |
| Depositing User: | Stralen Pratasik |
| Date Deposited: | 02 Feb 2026 03:07 |
| Last Modified: | 02 Feb 2026 03:07 |
| URI: | http://repository.its.ac.id/id/eprint/131494 |
Actions (login required)
 |
View Item |