Question

If for the first 29 days of a month between 10 and 16 people ride a certain city bus and then on the 30th day of the month 40 people ride the bus, should you consider the new data value (40 people in 1day) when predicting how many people will ride the bus each day next month? Why or why not?

Answer 1

No I will not consider this data value for predicting the number of people riding the bus each day next month.

If we plot the data in a plane, then we will see that all the points lie in a relatively small neighbourhood than this point. This type of data is called OUTLIERS in a given dataset, which differs significantly from the other data points.

In our case this could have for some special reason, there may be some public event at nearby place, Which is not the ideal case for normal days. If we take this data point in our prediction analysis, then this may cause overfitting to the predicted graph. So, the prediction have high chance of being wrong. So we avoid this data.