Regression Analysis

SOCIOECONOMIC IMPACT ON RATE OF LOW-BIRTH WEIGHT RATE PERCENTAGE

Effect Plot of All Predictors

The graph of HInc, PropHisp referred the negative relationship between rate and HInc, PropHisp as with the increase in Median Household Income and Proportion of Population of Hispanic Ethnicity, low birth weight rate decreased and the graph of PropBlack represented the positive relationship between PropBlack and rate as with the increase in proportion of Population of Black race, low birth weight rate increases. The slope of the regression line in graph of HInc, PropHisp, PropBlack was quite higher than other predictors, which represented the magnitude of the estimated effect of these predictors on the low-birth-weight rate response variable.

Hence, P value was less than 0.05 for HInc, PropBlack, PropHisp, it rejected the null hypothesis and concluded that there was significant relationship between these predictors supporting alternative hypothesis.

Residual Plots for all Predictors

The residual plot of HInc, Proprent, PropHism referred to the non-constant variance, but as it showed a slight curve, there needs to be some improvement in this model.

The residual plot of PropBlack followed by PopDen represented heteroscedasticity as the variance was unequal across the range value of the response variable of low birth weight rate in prediction.

Next Project

Deep Learning