The K-means clustering analysis for the wind speed in Kudat was then executed. The number of clusters was set to five. The initial cluster and final cluster of the wind speed in Kudat were obtained and shown in Table6. From observation, there are minor changes in the clusters' centre in between the initial process and final result. This was because the K-means clustering algorithm was finding the best mean value that could represent the entire cluster. For the prediction of wind speed in Kudat, the result for the final clusters' centre will be used as the best mean values for those five clusters. For 8 years of data, there will be a total of 96 numbers of wind speed data recorded.