ACO-based Projection Pursuit clustering algorithm

In order to find a more effective method of solving the problem of subjectivity and difficulty to deal with the high-dimension data in the clustering, a new method---an improved PP (Projection Pursuit) based on Ant Colony Optimization algorithm (ACO) was introduced. The ant colony optimization algorithm has the strong global optimization ability and the PP method is a powerful technique for extracting statistically significant features from high-dimension data for automatic target detection and classification. The ant colony optimization algorithm was employed to optimize the function of the projected indexes in the PP. Application results show that the method can complete the selection more objectivity and rationality with objective weight, high resolving power, and stable result.