The Evaluation And Comparison Of The Correlational Relation Levels Between Anaerobic Power And Spirometric Respiratoty Functions With Regression Analysis And Neural Networks In Young Individuals

Introduction
In this study, to find out whether there is a correlation between anaerobic power and respiratoric functions, regression
analysis (RA) and neural learning algorithm where subject to evaluation in Multilayer Perceptron (MLP).

Methods
157 male (20.29 ± 1.94) and 54 female (19.70 ± 2.17) - 211 volunteers in total, who had similar age, height, weight,
body fat (%), body mass index (BMI kg/m2 ) and who had a regular workout program for 4.0 ± 1.50 years - were
included in this study. Data was analyzed and compared in RA and MLP in the SSPS packed program.

Results
When male, female and general population were considered and calculated in both RA and MLP, significant positive
correlations were found between VC, FVC, FEV1, FEV1/FVC, PEF, FEF25-75 and MVV of spirometric respiratory
functions and anaerobic power values (p<0.05, p<0.01, p<0.001). The highest value of correlation (R) in RA for male
was FEF25-75 (R=0.341), in MLP, FEV1 / FVC (r=0.801), for female in RA, FEV1 and MVV (R=0.578), in MLP, VC
(r=0.623), and in total population in RA, FEV and MVV (R=0.657). The lowest correlation in RA in male was, VC
(R=0.190), in female FEV1 (r=0.311), in total population FEV1 / FVC (R=0.298), in MLP FEV1 (r=0.873). In addition,
average assumption error in RA and MLP calculations were, (RA=0.112, MLP=0.076) in male, (RA=0.109,
MLP=0.156) in female and (RA=0.125, MLP=0.059) in total population.

Discussion / Conclusions
When all the results are examined, it was observed that, MLP had reached a higher and stronger correlation (r)
calculation value and a lower error level than RA. These results revealed a positive high correlation between anaerobic
power and spirometric respiratory functions.

NOTA: O texto com a iconografia está no anexo

Arquivo