EXPERT SYSTEM SHELLS

Expert System Shells are generic systems that contain reasoning mechanisms but not the problem-specific knowledge. Early shells were cumbersome but still allowed the user to avoid having to completely program the system from scratch. The shell is really a ready-made Expert System without a knowledge base. All the programming components are there, waiting for rules to be entered into the system. Modern shells contain two primary modules: a rule set builder (to construct the initial knowledge base) and an inference engine (as the vehicle for arriving at conclusions) The basic structure of the Expert System Shell and its relationship to a knowledge database is as illustrated.

Expert Systems Shells Software Development Packages Exsys InstantTea K-Vision KnowledgePro

PATTERN RECOGNITION
INTRODUCTION "Pattern recognition is the research area that studies the operation and design of systems that recognize patterns in data. It encloses sub disciplines like discriminant analysis, feature extraction, error estimation, cluster analysis (together sometimes called statistical pattern recognition), grammatical inference and parsing (sometimes called syntactical pattern recognition). Important application areas are image analysis, character recognition, speech analysis, man and machine diagnostics, person identification and industrial inspection." Pattern recognition is a field within the area of machine learning. Alternatively, it can be defined as the act of taking in raw data and taking an action based on the category of the data. As such it is a collection of methods for supervised learning. Pattern recognition aims to classify data (patterns) based on either a priori knowledge or on statistical information extracted from the patterns. The patterns to be classified are usually groups of measurements or observations, defining points in an appropriate multidimensional space. A complete pattern recognition system consists of a sensor that gathers the observations to be classified or described; a feature extraction mechanism that computes numeric or symbolic information from the observations; and a classification or description scheme that does a actual job of classifying or describing observations, relying on the extracted features.

One of the most effective ways to fine the meaning in data is to find the patterns that represent it. In this way, the information can be understood by the computer as well as be easier to store with less memory requirements. The two major divisions of pattern recognition are vision and sound. The firsts goal is to get a computer to recognize pictures so that it can recognize objects in its surroundings that would be helpful in robotics. The latter wants to achieve a similar goal but is a primary concern with companies that want to produce a new means in which a person interacts with a computer by talking.

PATTEREN RECOGNITION AND AI:

The Centre prides itself on its international reputation and leadership role in the area of pattern recognition, the key technology of multimedia signal interpretation. Its research has contributed many of the state of the art methods in various aspects of pattern recognition system design. CVSSP techniques which help to identify the most discriminative pattern attributes, are used by practitioners worldwide. The Centre's expertise in classifier design spans statistical, decision tree, neural network and support vector machine approaches. The recent innovations made at the Centre, such as the predictive validation approach to modeling pattern classes help to solve as diverse problems as target and outlier (abnormality, fault) detection. The contextual decision making methods pioneered at the Centre have found applicability in 3D object recognition, aerial image matching, and land cover classification in remote sensing. The methodology of multiple expert fusion contributed by CVSSP has helped to enhance the performance of pattern recognition systems in many application areas. ACASVA is an example of one of the Centre's collaborative projects that combines pattern recognition, AI, audio-visual analysis and psychology.

RECOGNITION AND CLASSIFICATION PROCESS

The pattern recognition process

Step 1: stimuli produced by objects are perceived by sensory devices. The more prominent attributes produce the strongest stimuli.. The values of these variables are used to characterize an object in the form of a pattern vector X, as a string generated by some grammar, as a classification tree, a description graph, or some other means of representation. The range of characteristic attribute values is known as the measurement space M.

Step 2: A subset of attributes whose values provide cohesive object grouping or clustering, consistent with some goals associated with the object classifications are selected. Attributes selected are those which produce high interclass and low interclass grouping. This subset represent a reduction in the attribute space dimensionality and hence simplifies the classification process. The range of subset values is known as the feature space F.

Step 3: Using the selected attribute values, object or class characterization models are learned by forming generalized phototype descriptions, classification rules or decision functions. These models are stored for subsequent recognition. The range of the decision function values or classification rules is known as the decision space D.

Step 4: Recognition of familiar objects is achieved through application of the rules learned in Step 3 by comparison and matching of object features with the stored models. Refinements and adjustments are performed continually thereafter to improve the quality and speed of recognition.

APPROACHES FOR PATTERN RECOGNITION: There are two basic approaches to recognition. They are: Structured Description Statistical Description