Combining Supervised and Unsupervised Learning for GIS Classification

TL;DR

This paper introduces a hybrid learning algorithm combining fuzzy c-means and Minimerror for GIS classification, leveraging unlabeled data to improve accuracy, and compares it with traditional supervised methods.

Contribution

The paper proposes a novel hybrid incremental learning algorithm that integrates unsupervised and supervised techniques for GIS data classification.

Findings

The hybrid approach effectively utilizes unlabeled GIS data.

It achieves comparable or improved classification accuracy.

The method outperforms classical supervised classifiers in certain scenarios.

Abstract

This paper presents a new hybrid learning algorithm for unsupervised classification tasks. We combined Fuzzy c-means learning algorithm and a supervised version of Minimerror to develop a hybrid incremental strategy allowing unsupervised classifications. We applied this new approach to a real-world database in order to know if the information contained in unlabeled features of a Geographic Information System (GIS), allows to well classify it. Finally, we compared our results to a classical supervised classification obtained by a multilayer perceptron.