When to apply which discrete statistical distribution?

Statistics is an amazing subject which offers wide variety of statistical distributions to make predictions or to do forecasting. There are some discrete distributions which are defined for discrete random variables like number of telephone calls on a telephone exchange or number of children in a family. Some well celebrated discrete distributions are Binomial, Poisson, Geometric, Hypergeometric etcetera. On the other hand, we have continuous distributions like Normal, Beta, Gamma, Exponential etcetera which are defined for continuous random variables like lifetime of an electric bulb.

In this blog, my objective is to let…

We are standing in the last quarter of the year 2020. This year has been very difficult for the whole world. We did not know how to react to these weird situations. We did not know how to be efficient and productive in this pandemic. Something similar happened around 100 years back. I am referring to the Spanish Flu. So, we have not faced anything similar in our lives.Some were able to be productive in these unusual situations and some were not. And it’s completely okay because situation was (or is) very unusual.

But, now we have data of around…

Being a Statistician, i always admire the beauty of Statistics subject. Where it helps to forecast sales/profits/losses of an organization, it also helps to keep a check on mental health. Psychological Statistics helps to understand mental conditions and early detection of mental health related issues like depression. Predictive modelling can be used to predict mental health related issues. Psychological tests are designed in a technical way so that their score has a statistical significance as well. Statistics is just amazing!

Generalised linear models try to explain relationship between mean response and independent variables with the help of a mathematical function which is known as link function in literature. These models can be used to model any type of dependent variable - categorical and continuous both. These models offer wide variety of link functions like logit for logistic regression model, log link for Poisson regression model, probit for Gaussian regression etcetera. Here is some handy information:

Binary dependent: Logit link: Logistic Regression

Count data (dependent variable): Log link: Poisson Regression

If the above mentioned question popped up in your mind ever, this blog might help to answer this question. Mathematically, ANOVA might be considered as a special case of linear regression analysis in which all explanatory variables are categorical.

From the application point of view, there is a lot of difference between ANOVA and Regression. ANOVA is oftenly used to compare the dependent variable outcome against various categories of explanatory variables. Example - comparing sales of a product at different price categories. On the other hand, regression is used to predict the dependent variable with the help of explanatory variables…

Mathematical models explain functional relationship between y and x in non-random or non probabilistic environment. On the other hand, statistical models explain functional relationship between x and y in stochastic/ random/ probabilistic environment. The error term in a statistical model is responsible for introducing randomness.

Statistical model: y=b0+b1*X+e

Where y is response, X is a explanatory variable, e is the error term, bo and b1 are regression coefficients.

Data Scientist || M.Sc. Statistics (2018) || I like to express my thoughts through my writings.