Modelo predictivo para la valoración de la fuerza de agarre ajustado a una población de estudio
Date
2020-05Publisher
Pontificia Universidad Javeriana
Faculty
Facultad de Ingeniería
Program
Ingeniería Industrial
Obtained title
Ingeniero (a) Industrial
Type
Tesis/Trabajo de grado - Monografía - Pregrado
Share this record
Citación
Metadata
Show full item record
PDF documents
Abstract
The study of the physical attributes of the population has been an influential topic in recent years at the workplace, due to its
direct impact on the prevention of upper limbs' MSDs (Musculoskeletal Disorders). Currently, it is necessary to calculate grip
strength using a predictive model based on these attributes in the Colombian working population. The predictive model is
developed through a statistical analysis of a convenience sample (60 people). Forward and Stepwise methods are performed
on the training data that corresponds to a random sample (40 people) of the available data. Predictive equations scenarios for
grip strength, of both the dominant and non-dominant hands, are created and the data is divided by gender. The best scenarios
are then filtered according to the following criteria: absolute percentage error and p-value of the multiple linear regressions.
Four definitive equations are obtained and validated using the test data (20 people). These equations are introduced to an
application developed in Microsoft Excel, which has an interface developed in Visual Basic with the programming code for
the calculation of the grip strength according to the predictive model contained by the definitive equations, where (i) a person
inputs his/her information and obtains the result of their theoretical grip strength and (ii) any subject, interested in learning the
results of the study, can see a graphical representation of the prediction of forces for the test data and can compare his/her
calculated grip strength with his/her real grip strength.
To verify the design, the application was executed step by step, verifying that all its components and encodings followed the
established flow, and the results were shown automatically in the graphs and databases. Regarding the performance
requirements, the incidence of the predictor variables in the model was evaluated using the percentage error, minimizing the
differences and adjusting to the test data. Likewise, the application seeks to contribute to the prediction of grip strength in
subsequent studies and the field of occupational health in Colombia.
The app can predict the grip strength of a person and provide a comparison with other predictive models. Moreover, it has an
additional database illustrating the percentage errors calculated based on the real force and the obtained theoretical force. The
application aims to provide a tool to create a healthy environment for the working population, allows the calculation of
theoretical grip strength and it compares state of the art equations with the equation presented in this study and the real force.
Finally, the app asks the user to input the necessary information for the analysis of the strength of each individual, which is
stored in a database and provides the person or entity with the behavior of the theoretical grip force compared to the real force
so they can take the corresponding action.
Keywords
Fuerza de agarreModelo predictivo
Método Forward
Método Stepwise
Regresiones lineales
Ecuaciones
Themes
Ingeniería industrial - Tesis y disertaciones académicasMétodos de simulación
Métodos de simulación
Google Analytics Statistics
Collections
- Ingeniería Industrial [1112]