CLUSTERING OF NOISE DATA BY SELF-ORGANIZING MAPB. MAHALAKSHMI AND K. DURAISWAMY
Noise is an unwanted sound, intensity, or other feature that cause some kind of physiological or psychosomatic harm to humans or other living things. The sources of noise are more in urban and industrial areas than in rural areas. Modern life has given rise to this kind of pollution. Categorization has become important for handling large amount of data available for assessment. This paper proposes a method Self-organizing map for clustering of noise data. SOM is used for both clustering and projection. It projects onto a 2D-grid. Various methods were developed for the automatic clustering of data according to the user requirements. The objective of this paper is to reduce the time and effort the user has to spend to find the required information. The noise levels collected from various cities are tested using the methods 2D-SOM and 1D-SOM which produced better cluster results. The comparative results of 2D-SOM and 1D-SOM are given. The clusters formed reveal that cities with similar noise levels can be identified for assessment of noise pollution. The proposed method which uses 1D-SOM is found to be efficient in terms of computational time and clustering accuracy. It can be easily adapted for large data set.
Enter your contact information below to receive full paper.