Gabriel Fiol-Roig

A New Perspective in the Inductive Acquisition of Knowledge from Examples

In this paper we describe a new perspective in the Inductive Acquisition of Knowledge from Examples, based on three fundamental concepts: the Object Attribute Table (OAT), the Base of Attributes and Optimality Criteria and on a two step solution. The OAT constitutes an extensional description about some concepts to be intensionally described. To transform the knowledge from the OAT into an intensional form, the two step solution must be taken: n ni. n nTo obtain an optimal set of attributes or qualities to describe the concepts. n n n n nii. n nTo obtain an optimal intensional description based on the attributes obtained in the former step.

The Intelligent Butler: A Virtual Agent for Disabled and Elderly People Assistance

Social assistance constitutes an increasing problem in developed countries, which can be considered from two dimensions: the home and the hospital frameworks. Anyway, most of the tasks have to do with aiding people with limitations in complex environments as a hospital or a house. Intelligent agents constitute a powerful approach in designing computer systems making possible the interaction of the users (elderly and disabled people) with the elements of a domotics environment. Such a purpose can be achieved through the unification of artificial intelligence techniques, virtual reality, multimodal interfaces and digital nets with domotics services. This paper describes the design and implementation of the prototype for a virtual agent capable of attending disabled people in a home environment. The results are shown through a computer program that simulates the behaviour of the agent in developing some typical functions.

An algebra for the treatment of multivalued Information systems

Descriptive knowledge about an Information System can be expressed in declarative form by means of a binary Boolean based language. This paper presents a contribution to the study of an arbitrary multivalued Information System structure by introducing an algebra (not binary) that allows the treatment of multiple valued data tables with systematic algebraic techniques. Elements |ti| and ||tp||, called arrays and co-arrays, are defined, operations ~, ‡ and ∘ are described. The proposed methodology allows multivalued algebraic expressions describing a multivalued Information System (multivalued Object Attribute Table). Furthermore, the same Information System can be described by several distinct, but equivalent, algebraic expressions. Among these, the prime-ar expression is singled out. The usefulness of the described algebra to represent an Information System is shown.

UIB-IK: A Computer System for Decision Trees Induction

Decision Trees constitute a common knowledge structure to express the results of an Inductive Process. A computer system called UIB-IK, to induce decision trees from an initial collection of examples is presented. General properties of this tool are compared to those from some very known systems, such as the ID3, ID5, C4.5 and AQ11 systems. Performance qualities of UIB-IK are exposed on the basis of its functional model, and a synthesized description of two complex real applications is presented. The modular design together with the programming techniques used to implement the final program, makes UIB-IK to be a consistent and parameterized software tool, capable to cope a large range of problems.

Theoretical Considerations About Subset Descriptions

In this paper we present a new theoretical approach about intensional desciptions of subsets extensionally defined. In this approach some subsets of a certain domain are initially defined in an extensional way, in such a way that for each element of the subsets some attributes are known (the attributes are the same for each element). Obtaining an intensional description of the defined subsets consist of finding a property based on the attributes associated with the elements that allows classifying all the elements of the domain in their corresponding subset.

Data Mining Techniques for Web Page Classification

Nowadays, the Web is an essential tool for most people. Internet provides millions of web pages for each and every search term. The Internet is a powerful medium for communication between computers and accessing online documents but it is not a tool for locating or organizing information. Tools like search engines assist users in locating information. The amount of daily searches on the web is broad and the task of getting interesting and required results quickly becomes very difficult. The use of an automatic web page classifier can simplify the process by assisting the search engine in getting relevant results. The web pages can present different and varied information depending on the characteristics of its content. The uncontrolled nature of web content presents additional challenges to web page classification as compared to traditional text classification, but the interconnected nature of hypertext also provides features that can assist the process. This paper analyses the feasibility of an automatic web page classifier, proposes several classifiers and studies their precision. In this sense, Data Mining techniques are of great importance and will be used to construct the classifiers.

Applying Data Mining Techniques to Stock Market Analysis

The stock market can be viewed as a particular data mining and artificial intelligence problem. The movement in the stock exchange depends on capital gains and losses and most people consider the stock market erratic and unpredictable. However, patterns that allow the prediction of some movements can be found. Stock market analysis deals with the study of these patterns. It uses different techniques and strategies, mostly automatic that trigger buying and selling orders depending on different decision making algorithms. It can be considered as an intelligent treatment of past and present financial data in order to predict the stock market future behavior. Therefore it can be viewed as an artificial intelligence problem in the data mining field. This paper aims to study, construct and evaluate these investment strategies in order to predict future stock exchanges. Firstly, data mining techniques will be used to evaluate past stock prices and acquire useful knowledge through the calculation of some financial indicators. Next artificial intelligence strategies will be used to construct decision making trees.

Decision trees in stock market analysis: construction and validation

Data Mining techniques and Artificial Intelligence strategies can be used to solve problems in the stock market field. Most people consider the stock market erratic and unpredictable since the movement in the stock exchange depends on capital gains and losses. Nevertheless, patterns that allow the prediction of some movements can be found and studied. In this sense, stock market analysis uses different automatic techniques and strategies that trigger buying and selling orders depending on different decision making algorithms. In this paper different investment strategies that predict future stock exchanges are studied and evaluated. Firstly, data mining approaches are used to evaluate past stock prices and acquire useful knowledge through the calculation of financial indicators. Transformed data are then classified using decision trees obtained through the application of Artificial Intelligence strategies. Finally, the different decision trees are analyzed and evaluated, showing accuracy rates and emphasizing total profit associated to capital gains.

Formal model for real time diagnosis of dynamic systems

The task of real-time causal diagnosis of disturbances has been conceived traditionally from a procedural point of view, in the sense that the attention is focused on developing efficient procedures capable of evaluating the state of some variables of the system so that real time objectives imposed were satisfied. The main handicap of these methods lies in the difficulty to plan the diagnostic process, particularly when a high number of variables are to be observed model-based diagnosis constitutes a more complete approach to the topic. Considering the availability of a simulated model of the system, the task of the diagnostic procedure is now performed on the simulated model, facilitating the observation and handling of the variables of the system. However, the absence of languages allowing us to develop simulated models of real systems limits the use of this theory to simple cases. An approach to real-time causal diagnosis of dynamic systems based on a pre-established planning of any possible diagnostic situation in such a way real-time objectives are satisfied, is presented in this work. Artificial intelligence techniques, particularly inductive methods have been considered according to two essential steps: formulation of the causal diagnostic model, specifying the particular characteristics of the problem in hand; and generation of an information structure according to the characteristics of the formulated model, whose performance will guarantee the diagnostic objectives.

A Contribution for Elderly and Disabled Care Using Intelligent Approaches

This paper presents results obtained in an ongoing project that deals with home care assistance for the elderly and disabled. The problems faced in this project cover many disciplines and can be studied using different approaches. Nowadays, e-health constitutes a young and expanding area that uses new technological innovation methods for social assistance. Methods and techniques from the Artificial Intelligence field offer a broad range of ideas and points of view for solving the problem. In particular, information systems and intelligent agents are two perspectives that deserve further study. Information systems provide a formal knowledge representation that models important tasks such as concept description and decision making. On the other hand, intelligent agents provide a mechanism to implement a rational behavior. The combination of both perspectives offers a valid solution to our problems.