Step 0 : Machine Learning

In Machine Learning, Machine is a model or algorithm which defines the correlation between dependent and independent variable, That means it learns on correlation of the training data to make the future predictions.

In simple terms if a model can accurately identify if a bottle is red or green based on the previously training image set or if a customer is going to buy a bottle of ketchup if he bought some potatoes, then you prediction model is fitting the training set. Do dependent variable in later case is 'buying potatoes' and independent variable is 'buying a bottle of ketchup', it will be magical if we can predict accurately somehow that if customer will want a bottle so we can put it at a more approachable place. Sounds logical ?

All right before implementing the algorithm we assume you have a clean datasource. If you don't have clean dataset, you can find it over internet or clean it yourself using data cleaning methods.(find it on google)

So here is the questions to ask before choosing a machine learning algorithm. 

1. How is your input and output variables are related

2. Is your training set labeled or unlabeled

Well there are more parameters to consider, but majorly you can divide into two groups by asking above two questions, Supervised and Unsupervised Learning.

After you choose your algorithm, You can go on to apply the steps to make it a best fit or you can compare the result with results from another approaches.

In upcoming posts, we will dive deeper in algorithms.

Have fun with machine learning.

#algorithm #machine #machinelearning