<![CDATA[Malnutrition remains a major public health challenge in low and middle-income countries. This research proposes a method for diagnosing malnutrition diseases by combining Certainty Factor (CF) and Modified K-Nearest Neighbor (MKNN). CF is used to obtain certainty values from the symptoms experienced by patients, while MKNN classifies patient symptom data into specific disease classes based on proximity to the training data. The symptom CF values are combined with the rule CFs to obtain the final CF for each disease. The patient's CF data becomes the testing data, while the disease dataset is the training data. The MKNN technique involves calculating the Euclidean distance, validity, and applying weight voting to identify the class to which the testing data belongs based on the majority class of the k nearest training data. In the test case, CF indicated a tendency towards Kwashiorkor, reinforced by MKNN with the majority of the nearest training data classified as Kwashiorkor. Cross-validation testing with 20 testing data resulted in an accuracy of 95% for the combined CF-MKNN method. The combination of the two methods mutually reinforces and increases the confidence in the diagnosis.]]>
Copyrights © 2024
