Taguchi?s T-method with Normalization-Based Binary Bat Algorithm
Taguchi?s T-method (T-method) is a predictive modeling technique developed by Dr. Genichi Taguchi under the Mahalanobis-Taguchi system to predict unknown output or future state based on multivariable input variables. Conventionally, Taguchi?s orthogonal array is used as a variable selection approach...
Saved in:
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Conference paper |
| Published: |
Springer Science and Business Media Deutschland GmbH
2025
|
| Subjects: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Taguchi?s T-method (T-method) is a predictive modeling technique developed by Dr. Genichi Taguchi under the Mahalanobis-Taguchi system to predict unknown output or future state based on multivariable input variables. Conventionally, Taguchi?s orthogonal array is used as a variable selection approach in optimizing the predictive model. Due to a fixed and restricted predetermined design array, the orthogonal array is unable to give higher-order interactions between variables, resulting in inferior T-method prediction accuracy. Therefore, a variable selection technique using a swarm-based Binary Bat algorithm is proposed. Specifically, a normalization-based Binary Bat algorithm is used, where discretization of continuous solution into binary form is performed using a normalization equation. An experimental study was conducted, and the variable selection process using the normalization-based Binary Bat algorithm found a better combination of input variables which consists of only six out of eight variables. The prediction accuracy was also enhanced by 7.15% when validated using the validation dataset. In conclusion, the proposed method successfully yields better prediction accuracy as compared to conventional approaches. After the variable selection process using the proposed method, the optimal prediction model is now formulated with a lesser variable, making it less complex and computationally fast. ? The Author(s), under exclusive license to Springer Nature Switzerland AG 2024. |
|---|