Mobile App or Website?

AI can help you make correct decisions saving you costs while at the same time increasing your sales revenue. This article explains in simple terms how this is achieved using machine learning.

Introduction / Background

Regression analysis is a set of statistical methods used for the estimation of relationships between properties of an observation (subject). For instance if you are observing the growth of a child from infancy then you would like to check the relationship between the weight and height of that child. The child becomes the subject, the weight and height becomes the variables.

If you are studying how the weight changes with height then the weight becomes the dependent variable and the height becomes the independent variable. This means you can build a regression analysis that is able to predict the weight of a child given the height.

Linear regression analysis means there is a linear relationship between the dependent variable and the independent variable. Non-linear regression analysis means there is no linear relationship between the dependent and the independent variables. Meaning the relationship between the dependent variable and the independent variables is a complex one.

When studying how several independent variables affect a single dependent variable in a linear relationship then this is called a multiple linear regression. Using the same example, say you want to study how the weight changes with height and age, we will be having two independent variables affecting the one dependent variable which is the weight.

When studying how a single independent variable affect another single dependent variable in a linear relationship then this is called a simple linear regression problem and that will be our focus of entry in this blog.

In this article i will use a simple linear regression problem to explain how a business problem can be solved using simple linear regression. But just before i start, let me highlight some of the assumptions of linear regression:

1. Linearity of the residuals - The dependent and independent variables have a linear relationship between them
2. Independent of the residuals - The error terms should not be dependent on one another (like in time-series data wherein the next value is dependent on the previous one). There should be no correlation between the residual terms. The absence of this phenomenon is known as Autocorrelation.
3. Normal distribution of the residuals - The mean of residuals should follow a normal distribution with a mean equal to zero or close to zero. This is done in order to check whether the selected line is actually the line of best fit or not.
   If the error terms are non-normally distributed, suggests that there are a few unusual data points that must be studied closely to make a better model.
4. The equal variance of the residuals - The error terms must have constant variance. This phenomenon is known as Homoscedasticity.

Linear Regression

Using a hypothetical use case, we will assume XYZ Company dealing in cloths and cloth wear has both a mobile app and a website from which the customers place orders remotely. As the manager of this company you can use AI to arrive at a scientifically backed decision.

Business Problem

As the general manager of XYZ Company, I spend so much money in marketing cloths through mobile app and website but we are running low on budget. Where should i focus on the marketing campaign.

Data Available

Over the past one year, we have collected the below data for every customer:

Analysis

Because we are interested in sales by the customer, we shall now be interested to know how the other pieces of information (Email Address, Address, Avatar, Average Session Length, Time on App, Time on website, Length of Membership(Years) affect the Yearly Amount Spent)

By being able to predict, with a very high level of accuracy, the yearly amount spent then we shall have been able to get the formula that combines how each of the other pieces of information contribute to the yearly amount spent.

Technique Used for Prediction

This kind of problem involves determining the value of an output (what we call the prediction) as a function of the weighted sum of the input (these are the pieces of information made available and can be used to generate the output)

The technique used for this kind of prediction is known as Regression technique and the overall objective is to determine by how much each of the input information contributes to the yearly amount spent (output). Meaning if we choose Average Session, Length, Time on App, Time on website and Length of Membership (what we will call the input variables) as the information that is to be used to make the prediction (what we will call the output variable), then using Regression technique, we will be able to determine by how much each of the input variables contribute to a good prediction of the Yearly Amount Spent (output variable). These are what we call coefficients. So we are looking at identifying the coefficients of each of the input variables then we select the input variable with the greatest coefficient.

Types of Regression Techniques

1. Simple Linear Regression (Univariate)
2. Multiple Linear Regression (Multivariate)

Linear  Regression is a supervised learning algorithm in machine learning that supports find the linear correlation among variables. The result or output of the regression problem is a real or continuous value.

Other Real-world Examples of Regression

The below represents examples where regression technique may be used:

1. Forecasting sales : Organizations often use linear regression models to forecast future sales. This can be helpful for things like budgeting and planning. Algorithms such as Amazon’s item-to-item collaborative filtering are used to predict what customers will buy in the future based on their past purchase history.
2. Cash forecasting : Many businesses use linear regression to forecast how much cash they’ll have on hand in the future. This is important for things like managing expenses and ensuring that there is enough cash on hand to cover unexpected costs.
3. Analyzing survey data : Linear regression can also be used to analyze survey data. This can help businesses understand things like customer satisfaction and product preferences. For example, a company might use linear regression to figure out how likely people are to recommend their product to others.
4. Stock predictions : A lot of businesses use linear regression models to predict how stocks will perform in the future. This is done by analyzing past data on stock prices and trends to identify patterns.
5. Predicting consumer behavior : Businesses can use linear regression to predict things like how much a customer is likely to spend. Regression models can also be used to predict consumer behavior. This can be helpful for things like targeted marketing and product development. For example, Walmart uses linear regression to predict what products will be popular in different regions of the country.
6. Analysis of relationship between variables : Linear regression can also be used to identify relationships between different variables. For example, you could use linear regression to find out how temperature affects ice cream sales.

Sources

1. Vital Flux
2. Analytics Vidhya