Classifying fibre attributes of cotton using Kohonen neural networks

583-587Kohonen neural network has been used to classify cotton fibre characteristics, viz. 2.5% span length, bundle strength, short fibre index, uniformity ratio and elongation. Twenty three cottons studied are classified into 3 groups and each group is given range of values for each property. Except fibre elongation, the other properties have distinct range of values for each group. The developed model is found to have a classification rate of 100%, when validation is done for 4 cottons