This repo examines the performance of TSK fuzzy models in regression and classification tasks.
To explore the behaviours of fuzzy regression & fuzzy classification in various conditions, both are applied on two types of datasets:
- a simple dataset &
- a high-dimensional dataset
The project is structured following this rationale.
The implementation is in Matlab (R2019a) using the Fuzzy Logic Toolbox.
Datasets for Regression:
- Simple: Airfoil Self-Noise Dataset - https://archive.ics.uci.edu/dataset/291/airfoil+self+noise
- High-dimensional: Superconductivty Dataset - https://archive.ics.uci.edu/dataset/464/superconductivty+data
Datasets for Classification:
- Simple: Avila Dataset - https://archive.ics.uci.edu/dataset/459/avila
- High-dimensional: Isolet Dataset - https://archive.ics.uci.edu/dataset/54/isolet
All the above datasets are available through the UCI repository (https://archive.ics.uci.edu/)
| - 1.fuzzy_regression/
| - - airfoil_self_noise_dataset/
| - - - - airfoil_self_noise.dat
| - - - - src/
| - - - - - - Regression_TSK_model_1.m
| - - - - - - Regression_TSK_model_2.m
| - - - - - - Regression_TSK_model_3.m
| - - - - - - Regression_TSK_model_4.m
| - - - - - - split_scale.m
| - - - - - - plotMFs.m
| - - - - - - fis.fis
| - - - - evaluation_metrics_plots/
| - - superconduct_dataset/
| - - - - superconduct.csv
| - - - - src/
| - - - - - - plotErrorVsNumOfRules.m
| - - - - - - plotMFs.m
| - - - - - - Regression_Grid_Search.m
| - - - - - - split_80_20.m
| - - - - evaluation_metrics_plots/
| - - - - ranks/
| - 2.fuzzy_classification/
| - - avila_dataset/
| - - - - avila.txt
| - - - - src/
| - - - - - - Classification_TSK_classDependent_bigRadius.m
| - - - - - - Classification_TSK_classDependent_smallRadius.m
| - - - - - - Classification_TSK_classIndependent_bigRadius.m
| - - - - - - Classification_TSK_classIndependent_smallRadius.m
| - - - - - - plotMFs.m
| - - - - - - sug101.fis
| - - - - evaluation_metrics_plots/
| - - isolet_dataset/
| - - - - isolet.dat
| - - - - src/
| - - - - - - Classification_Grid_Search.m
| - - - - - - plotMFs.m
| - - - - - - split_80_20_Stratified.m
| - - - - - - sug191.fis
| - - - - evaluation_metrics_plots/
The data was normalized to the range [0, 1].