Question

What is the difference between a model and actual data?

Answer 1

Data are characteristics or information usually numerical that are used to depict the real life situation.It is usually a set of qualitative or quantitative variable about one or more object.

For extracting information from the data , the relationship between the variable is obtained which can be achieved by visualising the data.

Let us consider an example of data of average fall in India in year 2019. By looking at the data of rainfall, we can see that the highest rainfall occur in a month of September and lowest rainfall occur in month of May. This is just a piece of information which we can get by looking into the data .

Model-

It is usually a set of relationship between two or more random variable and some other variable. Model is generally of two types .

1. Mathematical model

In this case there is fixed relationship between the variable or we can easily determined fixed relationship between variable.

2. Statistical model

In this case the relationship between the variable is not completely fix. There is some random factor which occurs between the variable.

Statistical model is generally obtained by obtaining the relationship between the variable. Thus on the basis of relationship the statistical model is generally defined as of many type.

1. Linear model

There is linear relationship between the predictors and the dependent variable.

2. Exponential model

A model of the form

Represents exponential model with error

3. Curvilinear model

If there is no linear relationship between the variable.

Model is used for prediction or forecasting. Therefore it requires actual data and with the help of actual data the relationship can be easily obtained. Therefore the model between the dependent variable and two or more independent variable can be easily obtained.

Thus inshort we can say that model requires actual data to obtain the relationship. And thus the fitting of model can be tested by the deviation of the predicted data with the actual data.

The model with best fit has least deviation of predicted data from the actual data. This deviation can be observed by the root mean square deviation or residual sum of squares.