In the above files we gonna see about various timeseries tools to make a data set stationary for forecasting and many forecasting methods.
And all the topics discussed in the file will be discussed along with the practical examples .
- Time sereis is a forecasting method where the input data and the output data will be based on the time.
- It may appear that the time series looks similar to the linear regression model and you may also have a question on why there is a term seperately called time series when there is a term linear regression whiuch can do almost same thing as timeseries .
- If you take a deep look into both the topics then you will find the solution for your questions , Lets die deep into the topics :
-
Linear Regeression is Interapolation of data Whereas Timeseries is a Extrapolation of data.
-
Then what is meant by Interpolation nad Extrapolation ?
Lets take a example a very familiar one , A House Price Predcition model in which
step 1 :we provide the input and output for the training set and train the model
step 2 : Find the best fit line
step 3 : Then when the test set data is given the solution for the data can mostly be found with in the best fit line.
There is a chance of getting same values on both the axis that is a plot size and rate can be similar -
Whereas interms of extrapolation the valu of the x axis will be a timebased unit if we look little deeper a time is non repeated unit simply , we cannot have same date/month/year twice . once a time is crossed it cannot be reversed and the datas are based on time.
-
In Inteerpolation the forecasting will be within the predefined data given as input ,whereas in extrapolation the forecasting lies outside the datas given simply what will be price of gold tommorrow and how the price will be in next month is completely differnt from price of house with size of 1000 sq.ft
If you still have any douts regarding the topics we have discussed .Kindly ping me.
-