Statistical Validation Methods for Lactobacillus SP Growth: A Dental Application

Abstract

Background: Software for text analysis and data mining is called IBM SPSS Modeler. Predictive models were made using this software, along with other kinds of analysis. This work presented the multilayer perceptron analysis and simple linear regression analysis. Using this methodology, which finds the relationship between the objective variable and the predictor, the researchers may evaluate the usefulness of neural networks as a support tool for estimating the prevalence of the bacterium Lactobacillus Salivarius. Objectives: In this paper, we are going to determine the association between days of the culture and the rate of bacteria growth using simple logistic regression by improving it using multilayer perceptron. Methods: Simple Logistic Regression (SLR) and Multilayer Perceptron (MLP) were selected statistical tools for the factor determine as for the Lactobacillus. The significant analysis can be determined by SLR the next analysis is to validate the relationship between the dependent and independent variables using the MLP procedure. We tested the variables by evaluating the dataset using SLR by looking at the P value and R squared that showed the relationship between variables and the effectiveness of the dataset attributes. MLP model was created by using the holdout regression by partitioning the dataset into training and testing. This showed the mean absolute error and variance of the model, deeming it accurate due to its low error rate. The days of the culture and the rate of bacteria growth Lactobacillus are related each other in linear and exponentially. Conclusions: This study help researchers to understand the specific growth rate(s) which can be used to best grow the organism.