CURVE FITTING WITH POLYNOMIAL REGRESSION

In this project work, we look at how the least-square polynomial regression model is used to fit a non-linear relationship between a response variable and an explanatory variable in curve fitting. Finding a mathematical equation or model that best fits a noisy data has experience some drawback in curve fitting. i.e. finding an appropriate fit that best depicts the behaviour of the data. The purpose of this project is to show how the polynomial regression model can be used to show the relationship that exist between two variables; where the linear regression model is inadequate in describing such a relationship. The method of curve fitting used in this study is the least square polynomial regression method. It is designed in a way that, the model parameters are estimated by minimizing the residual term of the polynomial regression model; and then used the model to find the line that best fits the data points of the data set. This method was validated by modelling a data extracted from Nigeria Stock Exchange; and the model was able to predict over 80% of the relationship that existed in the data. It was discovered that the inadequacies of the simple linear regression model in describing the relationship that existed in a data set could be easily tackled by fitting a polynomial regression line. This is done by increasing the power of the independent variable to a higher power until we get a best fit.