| dc.description.abstract | This report concludes a year of investigative research in artificial intelligence. Artificial intelligence is a field of computer science in which recreating human behavior in machines is the main subject; the field is divided into four areas of research: pattern recognition, natural language, problem solving, and learning and reasoning. The first half of the year of research was spent learning about artificial intelligence in general, then the second half was spent focusing on pattern recognition.
First, pattern recognition was divided into two parts, feature extraction and feature recognition. Three algorithms were implemented on a large computer, the Isodata clustering algorithm, the K-mean clustering algorithm, and the Block-Nilson-Duda algorithm for feature extraction. The theory of these algorithms relies on analytic geometry and common quantities such as the average, the standard deviation, and the distance between two points.
Computers at the TAMU Cyclotron Institute were used to develop a set of subroutines, which are used in energy spectrum analysis, to extract peaks in the data and identify them. The subroutines follow the same procedure as would the human operator; they first approximate the background, subtract the approximation, and then look for peaks in the remaining data. The set of subroutines in appendix C of this paper is the result of the last year of research. | en |
