The Logical Rules of Commonsense Reasoning

The Logical Rules of Commonsense Reasoning

Xenia Naidenova (Military Medical Academy, Russia)
DOI: 10.4018/978-1-60566-810-9.ch004
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

In this chapter we describe a model of commonsense reasoning that has been acquired from our numerous investigations on the human reasoning modes used by experts for solving diagnostic problems in diverse areas such as pattern recognition of natural objects (rocks, ore deposits, types of trees, types of clouds etc.), analysis of multi-spectral information, image processing, interpretation of psychological testing data, medicine diagnosis and so on. The principal aspects of this model coincide with the rulebased inference mechanism that is embodied in the KADS system (Ericson, et al., 1992), (Gappa, & Poeck, 1992). More details related to our model of reasoning and its implementation can be found in (Naidenova, & Syrbu, 1984; Naidenova, & Polegaeva, 1985a; 1985b).
Chapter Preview
Top

Introduction

An expert’s rules are logical assertions that describe the knowledge of specialists about a problem domain. Our experience in knowledge elicitation from experts allows us to analyze the typical forms of assertions used by experts. As an example, we give the rules of an expert’s interpretation of data obtained with the use of Pathological Character Accentuation Inventory for Adolescents. This psycho-diagnostic method elaborated by Lichko (1983) is a classical example of an expert system.

Some examples of the expert’s rules are:

  • 1.

    “If (D - F) ≥ 4, then DISSIMULATION decreases the possibility to reveal any character accentuation and completely excludes the CYCLOID and CONFORM types of character”.

  • 2.

    “If the index Е > 4, then the CYCLOID and PSYCHASTENOID types are impossible”.

  • 3.

    “If the type of character is HYPERTHYMIA, then ACCENTUATION with psychopathies is observed in 75%, with transit disturbances – in 5%, and with stable adaptation – in 5% of all cases”.

  • 4.

    “If the index А > 6 and the index S > 7 and the index Con = 0 and the index D > 6, then the LABILE type is observed”.

  • 5.

    “If the index Е ≥ 6, then the SCHISOID and HYSTEROID types are observed frequently”.

  • 6.

    “If after the application of rules with the numbers x, y, z the values of at least two indices are greater than or equal to the minimal diagnostic threshold, then the mixed types are possible with the following consistent combinations of characters: Hyp - C, Hyp - N, Hyp - Hyst, C - L, L - А, L - S, and L - Hyst”.

We used the following abbreviations: Hyp - hyperthymia, C - cycloid, L - labile, А – asthenia, N – neurotic, S - schizoid, Con - conformable, Hyst - hysteroid, Sens - sensitive, D - dissimulation, F - frankness, Е - emancipation, and P - psychasthenia.

These rules describe the system of knowledge formed by the specialists through a long-time process of learning. One of inevitable and important steps in learning is connected with extracting objects or concepts from the observation data. The definition of object is a very complex task. Object is a phenomenon with invariable (to some extent) form and a set of constituent parts each of which can be described by a particular collection of attributes measurable or not measurable (qualitative). Object possesses a specifical way of genesis, it also possesses age and follows a certain law of development. The examples of natural objects are: forest, mushroom, berry, plant, tree, birch, fir-tree ect. Tree has crown, trunc, branches, leaves ect.

The specialists construct boundaries between objects in space and in time. By this reason, the following assertions appear: “with height ≥ 800 meters above see-level, there does not appear the meadow type of woodland”.

Object is connected with its environment. There exist, as a rule, factors determining the way of object development. The links between object and its environment factors are reflected by causal relations. For example, the forest is under the influence of landscape, climate, ground, and water conditions.

There are associative links between coexistent objects in time and space. For example, each type of woodland is associated with a certain type of predominent trees and vice versa. Predominant type of trees can have concomitant ones.

The properties of object can be independent or connected through causal (functional) links. So the age of tree determines the image and properties of crown, the age of forest determines the properties of its curtain. The images of forests on photographes are determined by the condition of aerial survey and the type of instrument.

In a forest region, the specialists study the distribution of the forest types. The boudaries between the types of forest are established and the sub-regions are generated. The number of sub-regions is generalized or harmonized. If some types of forest occur very rarely, then they are excluded from consideration, some types of forest can be joined in one type. Each type of forest (sub-region) is estimated with the use of a confidence coefficient, for example in terms of the frequency of its occurrence in the given region.

Complete Chapter List

Search this Book:
Reset