Halipova V. D., Malaychuk V. P., Tymoshenko Zh. I.
Oles Honchar Dnipropetrovsk National University
DETERMINING THE TYPE OF PERSONALITY ACCENTUATION USING FUZZY INFERENCE SYSTEM
Psychiatrists differentiate three health status of the person: the norm, accentuation, pathology. Accentuation of personality is the personality traits that tend to move to a pathological condition. Possible negative aspects of such state between normal and pathology usually occur under the influence of unfavorable external factors. There is the problem of determining the type of accentuation, especially by young professionals. For a survey of human real-time system it is required to support the decision about the type of accentuation. Currently, there are questionnaires (test Leonhard, Shmisheka) to determine the type of personality accentuation and basic observation by psychologist or psychiatrist. Tests in this case may give wrong information because person has some deviations from the norm and it is also assumed that the survey will lead by a young professional psychiatrist. Therefore it is being developed the decision support system in which the psychiatrist notes inherent personality symptoms and receives the response in the form of type (a combination of types) accentuation of personality.
K. Leongard identified 12 types of personality accentuation. Identity with only one type of accentuation is rare. Usually, the person with borderline mental health has several accentuations. One can have the most obvious combination, as demonstratively and stuck personality, and quite strange at the first glance compound introversion and hypomania. In the other case, it can be easily confused extroverted and hypomanic personality.
Accentuations have been analyzed and found 80 symptoms. Descriptions of accentuation types have been made by K. Leongard. Among them there are symptoms which are typical for multiple accentuations or only one of them.
Symptoms of accentuations that inherent personality, can be represented in the form of linguistic variables. They differ from the numeric variables so that their values are not numbers, but words or sentences in natural or formal languages. Since words are generally less accurate than the numbers, the concept of linguistic variable enables approximately to describe the phenomena that are so complex that they cannot be described in conventional quantitative terms. Person’s mental state corresponds exactly to that description.
Each symptom is corresponded by a rank, the maximum number – if strongly pronounced, and so, in descending order to rank 1. All possible ranks form the base set of the linguistic variable. Thus, the symptom will be characterized by the term-set membership function with maxima in a given by psychiatrist number rank.
Fuzzy inference is an approximation of depending «input-output» on the basis of linguistic utterances «If – then» and logical operations on fuzzy sets. Fuzzy inference is used for modeling objects with continuous and discrete output. Objects with discrete output adequate for classification of medical and technical diagnostics and decision-making in other areas. The developing decision support system of the type of personality accentuation is also the system with a discrete output.
Definition of accentuation of personality in the theory of fuzzy sets is fuzzy inference for classification tasks. Classification problem consists of assigning an object given by the informative signs vector to one of the classes described in advance.
In this case, influencing parameters are the symptoms of accentuation that have personality. Class name is given as a part of a term-set output variable. In this case a fuzzy inference Sugeno is applied. Multidimensional depending inputs and outputs define fuzzy rules with logical operators AND and OR.
Definition of accentuation of personality is an important direction for timely diagnosis and helps the person’s state of mind not turn into a pathological. For this purpose, proposed a model with description of symptoms accentuation as linguistic variables and further processing of information by classification on the basis of fuzzy inference. The proposed mathematical model will be implemented in computer technologies of the support decision about the type of personality accentuation.