The learning system LERS (Learning from Examples based on Rough Sets) induces a set of rules from examples and classifies new examples using the set of rules induced previously by LERS. These rules are more general than information contained in the original input data since more new examples may be correctly classified by rules than may be matched with examples from the original data.

LERS handles inconsistencies using rough set theory. The main advantage of rough set theory, introduced by Z. Pawlak in 1982, is that it does not need any preliminary or additional information about data (like probability in probability theory, grade of membership in fuzzy set theory, etc.). In rough set theory approach inconsistencies are not removed from consideration. Instead, lower and upper approximations of the concept are computed. On the basis of these approximations, LERS computes two corresponding sets of rules: certain and possible.

The first implementation of LERS was done by John S. Dean and Douglas J. Sikora in 1988. Other important steps were: adding module LEM1, module LEM2 (with Chien - Chung Chan, as a part of his Ph. D. research), and improvements in the basic algorithm (with Sachin Mithal), improvements in implementation (Lei Yue), and the fundamental implementation by Michal R. Chmielewski, Paolo Werebrouck, Alfian Budihardjo and Gabriel Gonzalez. The system was improved, among others, by Soe Than, Chien - Pei B. Wang, Andi Yan, Xihong Zou, and Xinqun Zheng.

The machine learning/ data mining system LERS is universal. It may induce rules from any kind of data. One of potential applications is the use of expert systems, equipped with rules induced by LERS, as advisory systems, helping in decision making and improvement strategy.

LERS has potential applications to

The machine learning/ data mining system LERS has proven its applicability having been used for years by NASA Johnson Space Center (Automation and Robotics Division), as a tool to develop expert systems of the type most likely to be used in medical decision - making on board the International Space Station. LERS was also used to enhance facility compliance under Sections 311, 312, and 313 of Title III, the Emergency Planning and Community Right to Know. The project was funded by the U. S. Environmental Protection Agency. LERS was used in other areas as well, e.g., in the medical field to assess preterm labor risk for pregnant women and to compare the effects of warming devices for postoperative patients. Currently used traditional methods to assess preterm labor risk have positive predictive value (the ratio of all true positives to the sum of all true positives and false positives) between 17 and 38%, while the expert systems with the rule sets induced by LERS have positive predictive value between 59 and 93%.

Return to the Main Page