Question

1. Which ONE of the following are regression tasks? *
Predict the age of a person
Predict the country from where the person comes from
Predict whether the price of petroleum will increase tomorrow
Predict whether a document is related to science

Answer 1

Regression

Regression is just a mathematical technique to analyze relationships between some dependent and independent variables.

Simply speaking, regression takes in some known data about something, and fits in a mathematical curve or model, using which we can predict future data about the same thing with some level of confidence.

For example, consider the relation of Coronavirus Cases and the Population Density. Here, the population density is the independent variable. The number of cases is the dependent variable (as it depends on the population density).

We have a prior set of data from some places. We apply the regression techniques and fit a model to the known data. Now, using this, we can predict what will be the number of Coronavirus cases in some different area using its population density. It may not be exact, but it comes from a mathematical model using some prior data.

$\rule{300}{1}$

Let's look at the current options.

1) Predict the age of a person

There is just one variable. The age of the person. How are you going to predict it? There's no prior data. Not a regression problem.

2) Predict the country from where the person comes from

The same problem as the first one. We don't have anything to fit a model to. No data. No other variables. Not a regression problem.

3) Predict whether the price of petroleum will increase tomorrow

This is a Regression Problem. We can collect the data of the price of petroleum over the many days and we can immediately see a general trend. Whether the prices are increasing or decreasing. We can fit a linear curve on Price of Petroleum vs Date and get a model to predict tomorrow's price.

4) Predict whether a document is related to Science

This is more of a Classification Problem. This sort of faces similar issues with the first and second options. Not a regression problem.

Technically, we could fit a mathematical model by creating variables like words, vocabulary, math involvement, technicality and all. But, it is somewhat far fetched and we already have our answer as the third option.

$\rule{300}{1}$

The Answer

3) Predict whether the price of petroleum will increase tomorrow