Data mining refers to the preset procedures and algorithms used to extract these valuable patterns. The research aims to improve partition-based clustering algorithmswith advanced features of efficient data analysis and automatically generate an appropriate number of clusters.The efficiency of K-Means clustering is further challenged by real-world datasets with high dimensionality. As a result, the algorithm becomes too expensive to implement. With an increase in size comes a decrease in cluster quality. This study proposes a K-Modes algorithm-based technique for working efficiently with large dimension datasets.Improvements can be made to this approach by eliminating non-significant features from the clustering process, which reduces the dimensionality of the clusters created and improves their accuracy. However, this number can be used as an input depending on user requirements if it has a significance value greater than or equal to 60% of the maximum significance value in the proposed algorithm.

Data Mining, K-Means, Clustering, IoT