Predicting Gender Based on Severity of Symptoms of Schizophrenia and Cognitive Parameters

Predicting Gender Based on Severity of Symptoms of Schizophrenia and Cognitive Parameters

Kadir Uludag
Copyright: © 2024 |Pages: 12
DOI: 10.4018/979-8-3693-2426-4.ch014
OnDemand:
(Individual Chapters)
Available
$37.50
Current Special Offers
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

Abstract Schizophrenia (SCZ) is a complex and chronic psychotic disorder characterized by a range of symptoms that affect the psychological well-being and stability of patients. These symptoms can be categorized into positive, negative, general, and cognitive domains. Goal:Our objective was to utilize a combination of symptoms exhibited by patients with SCZ and various cognitive variables to predict their gender. Methods: We recruited a total of twenty-three patients diagnosed with SCZ for our study. Results: Based on the primary findings of our study, the Random Forest (RF) model demonstrated a remarkable accuracy of 85.71% in predicting gender, with a sensitivity of 50% and specificity of 100%. Additionally, the Neural Networks (NN) model achieved an accuracy of 87.5% in the training set and 28.5% in the test set for gender prediction, with a sensitivity of 0% and specificity of 50%. Conversely, the Logistic Regression (LR) model exhibited lower performance, with an accuracy of 42.85%, sensitivity of 0%, and specificity of 60% in predicting gender.
Chapter Preview
Top

2. Methods

2.1. Participants

For this study used data taken from a previously published study related to working memory and schizophrenia (SCZ) (Repovs and Barch, 2012). The persons with SCZ were outpatients and were on antipsychotics for a period of a minimum of 2 weeks.

The participants for this study were engaged through the Conte Center for the Neuroscience of Mental Disorders at Washington University. Each participants were cured with DSM-IV criteria of SCZ (n= 23).

Table 1 includes symptom severity scores according to gender, while Table 2 includes accuracy scores according to three different models. Tables 3, 4, and 5 also comprise confusion matrixes that allow calculating accuracy scores.

Most of the participants were male and were recorded as having a white race (n=14).

All subjects were diagnosed according to an agreement between a research psychiatrist and a research assistant who used the Structured Clinical Interview for DSM-IV. Each indivuduals were not included if they: (1) met DSM-IV criteria for substance dependence; (2) had an unstable medical disorder, or (c) head injury. All participants submitted written consent for participation in the study.

2.2. Statistical Analysis

All analyses were done in Python version 3.8.8. Significance levels are for two-tailed tests, and P < 0.05 is considered a significant value.

The t-test for independent samples or the Mann–Whitney U test was used to compare parameters between the gender groups. There was no missing variable to report.

2.3. Random Forest

Random forest (RF) (Breiman, 2001) is a well-functioning machine learning (ML) algorithm that is commonly used to solve classification problems in many scientific fields, such as geography, genetics, psychology, and psychiatry.

2.3.1. Random Forest Parameters

Maximum depth was set to “2” and the rest of the parameters were default values of scikit-learn (Pedregosa et al., 2011).

Complete Chapter List

Search this Book:
Reset