Introduction to Machine learning for managers

Introduction to Machine learning for managers

Before we explore the world of machine learning, we need to understand a few basics of Machine learning.

This will bring you in the comfort zone as many managers lack engineering knowledge.

Input:

The raw data required for the analysis. This data may be from survey, observations, or from data sets. 

Marketing: Customers purchased data in a Big bazaar retail out let for a month..

Human Resources: Employees training data for last ten years.

Finance: Stock price of FMCG products in last five years on a daily basis.

Operations: The cost of packaging for multiple orders from multiple customers in the firm for last two years.

Algorithm:

Machine learning field works on the algorithms to identify the predicted value. Thus, it is necessary for us to know basics of algorithms. These are set of rules a manager introduces to get the forecast value.

Let us understand with some examples;

Marketing: 

A manager would like to know the best sales promotion tool for his online marketing campaign in different markets. The set of rules are as follows:

Step1; START

Step2; Define promotional too used by a company Search ads, display ads,

Step 3; Define different markets: Panaji, Mumbai, Bengaluru, Delhi, and Chennai.

Step4; Define revnue from diffrent markets.

Step 5: Define cost of promotion types in different markets.

Step 6; Calculate Revenue per market per campaign.

Step 7: Calculate the best ROI for a promotion on a market wise

Step 8; STOP.

Human  Resource Management:

A human resource manager would like to know influence of educational qualification in completing a particular project. This will help them in recruiting a better candidate next time. The rules can be;

Step1: START

Step 2: Extract the education qualification data from employee records say CS, IS, Mechanical, Civil etc...

Step 3; Collect project time taken by each employee.

Step 4: Identify the project difficulty level easy, medium and difficult.

Step 5: identify the best combination of difficulty, qualification and optimum time.

Step 6: STOP.

Finance:

A finance manager want to construct a portfolio of stocks to recommend his clients. The algorithm can be set as below:

Step 1; START

Step 2 : Collect P/E ratio of all NSE stocks and set range of expectation.

Step 3: Collect EPS data for all NSE stocks

Step 4: Collect dividend announced by companies in last five years.

Step5: Define sector names ( Automobile, Bank, IT etc..0

Step 6: Calculate top 10 companies a sector wise on P/E, EPS and Dividend.

Step 7; Collect predicted growth rate for each sector.

Step 8; Divide company in Step 6 based on STEP 7 paramter.

Step 9; STOP