Dendral And Mycin Are The Best Expert Systems Ever Build

Expert system is a computer application that focuses on stimulating the intelligence of human expert in decision making. Every expert system mainly comprises of three parts – the inference engine, the knowledge base and the user interface. Based on these three parts the expert system infers new rules, adds them to the knowledge base, if not already present, and finally takes decisions based on the knowledge acquired in the knowledge base. Expert systems are great discoveries in the field of artificial intelligence. Out of the several expert systems available today, we would be discussing two most common and useful expert systems.

Dendral

Dendral is the expert system that is developed for structure elucidation of chemical compounds. This project began at Stanford University in 1965. Dendral is mainly concerned with the interpretation of data that is obtained from a device called mass spectrometer. In order to perform the structure elucidation of chemical compounds the expert system need to have knowledge about that specific domain, that is, the knowledge needed for reading and interpreting the output of various devices like infrared spectroscopy, ultraviolet spectroscopy etc.

The working process of dendral follows “plan generate and test” policy. That is, a search process is initiated which searches the states and then test each search state and execute them. In dendral, instead of generating every single possibility, some planning is done and only a single possibility is generated. Constraint Generator (CONGEN) is a dendral program that finally constructs the complete chemical structure by first taking the molecular formula as input and then generating all the possible chemical structures based on some constraints.

Mycin

Mycin is one of the oldest and successful expert system that is used to diagnose bacterial infectious diseases. It was also developed in the Stanford University in the year 1976. Its main aim is to help physicians to identify which bacteria has been the cause of the infection and suggest appropriate remedies to cure it. The basic organization of mycin consists of the following components:

Consultation program: This program interacts with the user by asking them questions about a particular case based on patient’s data and previous gathered knowledge. Based on this information catered from the users mycin infers information and gives advice accordingly.

Dynamic patient data: This contains all the information about the patient.

Explanation program: This part never gives any answer on its own, it only justifies the advice given by the expert system using static knowledge.

Static knowledge base: This is the static area where the expert system stores all the knowledge acquired from previous experiences and refers to this knowledge base whenever any similar case in confronted.

Knowledge acquisition program: This is the program that is concerned with the knowledge acquisition and storing them in the knowledge base and changing the data whenever required and also adding new rules to the knowledge base.

In mycin to gather input data, some samples are taken from the site of infection. Then these samples are sent to microbiology laboratory for further tests. Mycin was basically developed to assist a physician, who is not an expert in the field of antibiotics for the treatment of blood infections.