<![CDATA[Modified K-Nearest Neighbor (MK-NN) has been widely used to classify various types of objects. In carrying out the classification, MK-NN calculates the distance k closest neighbors in the training data. The difference between K-Nearest Neighbor (K-NN) and M-KNN is found in the process of calculating the validity of all training data and weight voting. The MK-NN algorithm calculation stage is calculating the distance between training data, calculating the value of the training data validity, calculating the distance between the training data and test data, and calculating weight voting. The biggest weight voting results taken are the number of K used. From the weighted voting results, the class of the largest weight voting value is the disease class from the test data. Potato plant data (Solanum tuberosum L) were used as many as 115 training data and test data with 7 types of diseases and 23 disease symptoms. The accuracy of this system depends on the k value and the total training data used. Big value of K make small the accuracy because the validity value obtained is smaller. The more training data used, the higher the accuracy because the difference between Euclidian grades between classes is greater. The best system accuracy is obtained from the value of k = 4 and total training data of 45 is 97.142857%.]]>
Copyrights © 2021
