regression with lagged values data analysis tool in excel

If the regression model is unable to adequately capture the ‘information’ contained within these correlations, the ‘unexplained’ information will leak into the model’s residual errors in the form of auto-correlated errors. Previous counts influence the values of future counts. If the data set is a time series of counts, additional modeling complications arise because time series data are often auto-correlated. Ordinary Least Squares Regression based linear models or non-linear models such as those based on Neural Nets based regression techniques don’t work well for such data sets because they can predict negative values. For example, the number of people walking into the emergency room of a hospital every hour is one such data set. Poisson and Poisson-like regression models are often used for counts based data sets, namely data that contain whole numbered counts.

How to build a Poisson regression model for data sets that are a time series of counts, using Python and Statsmodels

Regression with lagged values data analysis tool in excel