Download PDFOpen PDF in browserComparison of Different Cluttering Validity Methods in the Evaluation of Results for Finding Electrical Fault Location in Industrial MV Network Using Fuzzy Clustering TechniqueEasyChair Preprint 2275, version 55 pages•Date: September 3, 2020AbstractComprehensive evaluation is carried out to compare three different Cluttering Validity methods; Partition Coefficient, Partition Entropy and Proportional Exponent in the Evaluation of the results for Finding Electrical faults in industrial MV network using fuzzy clustering technique. Different data normalization methods and different range of alfa cut values for defuzzification are considered in the comparison process. The result shows that using Partition Entropy with Maximum Matrix normalization and 50% alpha cut gives the best effort saving of 95.15 in finding the fault Keyphrases: Cluster ValidityFinding Fault, Network Distribution, fuzzy clustering
|