Downloads: 164 | Views: 181 | Weekly Hits: ⮙1 | Monthly Hits: ⮙1
Research Paper | Computers in Biology and Medicine | Kenya | Volume 8 Issue 5, May 2019
Inter-County Comparative Analysis of ID3 Decision Tree Algorithms for Disease Symptom Burden Classification and Diagnosis
Nicodemus Nzoka Maingi | Ismail Ateya Lukandu | Matilu Mwau
Abstract: The ID3 decision tree algorithm provides a key method of defining decision trees that can be used to prioritize and eventually classify disease outbreak symptom burdens in the fight against disease outbreaks. The decision trees are mainly derived from the calculation of the entropy of using some predefined variables of interest, herein referred to as disease symptom burden variables (which generally point to any diseases symptoms coded into variables accordingly) and then ranking of information gain ratios of the various disease symptom burdens. The decision trees can then be compared to draw important ideas and knowledge. The comparison of the decision trees for various geographical regions (counties) provides key ideas to better understanding of the various similarities and differences, be they just pure random, geographical, or even deliberate. This comparative understanding can help the relevant authorities in better joint policy development and business continuity planning in the event of any disruptive disease outbreaks. The comparison could trigger some critical vantage points; providing better economies of scale in running joint surveillance activities as compared to individualized planning and executions, pooling efforts together to create useful and unassailable synergistic styles of execution, and finally it also allows the various teams bring in unique skills and experiences that wouldnt have been possible in separately executed endeavors. Ultimately, such efforts could also help the health and government personnel get to easily identify common attributes and results that could prove key in fighting disease outbreaks. Since the algorithm used here breaks down each disease into its constituent symptomatic burdens, it helps to cluster together those attributes or symptom burden variables that are most critical in the fight against disease outbreaks instead of the traditional focus on the general diseases alone.
Keywords: Trees, Entropy, Information gain, ID3, Disease Symptom Burdens
Edition: Volume 8 Issue 5, May 2019,
Pages: 83 - 89