Artificial Intelligence in Medicine

Expert Systems

Is today the area of application of AI within medicine most successful. Expert systems can store and use knowledge of one or more human experts in a particular application domain. Its use increases productivity, improves efficiency in making decisions or simply de-solves problems when experts are not present. There are many examples of developed expert systems.

method


The operation of Mycin was mainly based on simple inference engine, which drove a knowledge base of approximately 500 rules. The program captured the entrances from a series of questions (such as, Does the patient have chest discomfort or ?, Was the patient previously operated?), Which usually responded the patient's physician.

After this process, Mycin showed the exit screen, which consisted of a number of possible diseases (ordered by its associated probability), the explanation of why each of these diagnoses, and a series of recommendations on the treatment to be followed by the pacient. To calculate the probability of each of the results, the authors developed an empirical technique based on certain factors.

These certainty factors were calculated such that according to some evidence to hypothesis assigns a certainty factor. 

Results

Research conducted by the Stanford Medical School, revealed that Mycin had a hit rate of about 65%, which improved statistics Most doctors do not specialize in the diagnosis of bacterial infections (domain in which Mycin was specialized ), exercising the profession in those years. Doctors working exclusively in this field getting a rate of 80%.