close this bookVolume 3: No. 22
View the documentGovernment/military news
View the documentGovernment grant news
View the documentComputer industry ventures
View the documentComputational news
View the documentNeural-network reports
View the documentData sources
View the documentJob opportunities
View the documentJournals and journal calls
View the documentGames
View the documentComputational linguistics
View the documentTravel tips

Jack Ganssle ( is starting an algorithms column for hardware and assembly language programmers. His first column included "A OR B == ((NOT B) AND A) + B", "A XOR B == A + B - ((A AND B) << 1)", and the binary-bit-to-Gray- code conversions "G == B XOR (B/2)" and "B(i) == G(i) XOR B(i+1)". [EDN, 1/28/93, p. 25.] (Is there some standard compilation?)

The Software Patent Institute (SPI; Ann Arbor, MI) is seeking contributions to a folklore database with free-text descriptions of computational tricks and techniques. The database will be made available to the public in a few months. Submissions can be via email or PC/Mac diskette. Contact, (313) 769-4064 Fax. [Bernard A. Galler, CRN, 5/93, p. 10.]

For a comparison of nine inductive machine-learning algorithms on 130K cases, FTP report from /pub/TechReports on It's "An examination of inductive learning algorithms for the classification of sleep signals," by John A. Bentrup ( and Sylvian R. Ray. Overall winners for high accuracy with simple rules were C4, MDL, and AIMS. PRG was significantly more accurate when able to classify events, but at the cost of a high rejection rate. COBWEB and nearest-neighbor classification have advantages when the amount of training data is limited. [, 5/29/93.] Another paper is M.W. Kattan, D.A. Adams, and M.S. Parks, "A Comparison of Machine Learning with Human Judgment," J. of Management Information Systems, (9,4), Spring 1993, pp. 37-57. [Mike Kattan (]

Fuzzy Logic Applications Group (FLAG) is a non-profit users group in Vancouver. A newsletter and monthly meetings are planned, starting 6/3. John Meech (, (604) 822-3984, (604) 822-5599 Fax. [, 5/26/93.]