Density-based data clustering method

A density-based data clustering method executed by a computer system, comprising:

• a setup step setting a radius and a threshold value;
• a clustering step defining a single cluster on a plurality of data points of a data set, and providing and adding a plurality of first boundary marks to a seed list as seeds;
• an expansion step expanding the cluster from each seed of the seed list, and providing and adding at least one second boundary mark to the seed list as seeds;
• and
• a termination step determining whether each of the data points is clustered, wherein the clustering step is re-performed if the determination is negative,
• wherein the expansion step comprises:
• a first retrieval step retrieving one of the seeds in the seed list, defining another searching range for the retrieved seed based on the radius, and setting the data points located in the searching range of the retrieved seed as neighboring data points;
• a first determination step determining whether a number of the neighboring data points in the searching range of the retrieved seed is larger than or equal to the threshold value, wherein the neighboring data points in the searching range of the retrieved seed and the initial data point are clustered together as the cluster, and the at least one second boundary mark is added to the seed list if the determination of the third determination step is positive, and the first retrieval step is re-performed if the determination of the third determination step is negative;
• and
• a second determination step determining whether each seed in the seed list is retrieved, wherein the first retrieval step is re-performed if the determination of the fourth determination step is negative.