Local Transfer Function Classifier (LTF-C) is an artificial neural network for classification problems. It has similar architecture as Radial Basis Function (RBF) network, but employes entirely different training algorithm. LTF-C performs very well in real-world problems, like: handwritten digit recognition, credit risk assessment, classification of breast cancer tissue.
To learn more about LTF-C, see:
- Paper with detailed description of LTF-C (pdf)
- Report (pdf) for the 2nd EUNITE Competition - details of application of LTF-C to Modeling the Bank's Client Behavior (LTF-C won this competition!)
- LTF-Cimulator - Windows application for simulating LTF-C