A modified dynamic evolving neural-fuzzy approach to modeling customer satisfaction for affective design

Area of Science:

• Engineering
• Computer Science
• Human-Computer Interaction

Background:

• Affective design is crucial for product development and competitive advantage.
• Neural-fuzzy networks effectively model customer satisfaction, handling fuzziness and non-linearity.
• Existing neural-fuzzy models have limitations with numerous inputs and adaptability to new data.

Purpose of the Study:

• To propose a modified dynamic evolving neural-fuzzy approach to address limitations of conventional methods.
• To enhance the modeling of customer satisfaction in affective design.
• To improve adaptability and scalability for complex product development scenarios.

Main Methods:

• A modified dynamic evolving neural-fuzzy approach was developed.
• A case study focused on the affective design of mobile phones.
• Validation tests compared the proposed model against Adaptive Neuro-Fuzzy Inference System (ANFIS) variants.

Main Results:

• The conventional Adaptive Neuro-Fuzzy Inference System (ANFIS) failed due to a large number of inputs.
• The proposed dynamic neural-fuzzy model demonstrated superior modeling accuracy.
• The proposed model showed better computational efficiency compared to subtractive clustering-based and fuzzy c-means clustering-based ANFIS models.

Conclusions:

• The modified dynamic evolving neural-fuzzy approach effectively models customer satisfaction in affective design.
• The proposed method overcomes the limitations of traditional ANFIS, particularly with large input datasets.
• This approach offers enhanced accuracy and computational efficiency for product development.