UGA: A New Genetic Algorithm-Based Classification Method for Uncertain Data
Ava Assadi, Saman Harati Zade · 2014
Correct diagnosis of disease, is always one of human problems. Nowadays with advancement in the field of computer science, computer solutions can be used to solve this problem. One of these solutions is to use machine learning and data mining to detect disease. In this area, so many works have been done so far, but most of them, assumed data to be certain, whereas, in the medical field, the probability of data to be uncertain, is so much more than the other fields. Data uncertainty is common in real-world applications due to various causes, including imprecise measurement, network latency, out dated sources and privacy. To achieve this goal, there are many techniques, but physicians are very interested in rule based techniques that extract if - then rules, because these techniques are very easy to understand and also the conclusion is clear. So we decided to present a rule-based algorithm to diagnose diseases. We used genetic algorithm to extract fuzzy classification rules from data. In our suggested algorithm, we tried to improve the accuracy. Our experimental results show that our suggested algorithm has better performance than the other rule based related works.