What is Machine Learning? For Ultimate Beginners..
Hey everyone.. Welcome to my first story on Medium..
What can you expect from this story? Well.. It is not about any algorithm, it is a simple real life incident which I would be sharing with you to understand on what is Machine Learning..
It’s a simple story how my brain got trained and it took a decision with respect to the observed values and made the calculation..
It is always fun to learn any subject by relating the subject matter to a real life scenario. Since the day I came home to my parent’s place for vacation, I have been observing the quantity of milk that we have been getting every day. We buy milk in two ways, first- the traditional Indian way of getting it from the milk man, who comes all the way to home and the another way - my dad would get it on his way back from office.
The observation was on the quantity of milk purchased every day. Let’s make the explanation slightly technical. Just have a look at the table below
A small description on the table: The days in orange depicts week-1 and the purple for week-2. You can notice the last row of the table where the quantity in Method-1 was to be calculated with the known value of Method-2.
On Thursday of 2nd week, my father was on leave to office, which obviously makes the value to be zero (0) in the Method-2. The problem aroused when I went to collect the milk from the milk man. The values are well tabulated now, i did not have it when I had to make the decision. It was done on approximation. But consider the same scenario in terms of a machine. What if a machine needs to make the decision in such a situation?
Observations from Table:
- The values are uncertain and there is no common or rational relationship between the values(quantity) for each day.
- The value of Method-2 is zero(0) in the weekend (Saturday & Sunday).
- To balance with the stock in the weekend, the value of Method-2 for Friday is slightly high comparatively. A precaution is taken.
It was uncertain for the system to predict the value for method-1 on a week day which hasn’t occurred in the recent observation. For the system to calculate the value, it should go through a small analysis with the set of data.
How would the system calculate?
Was there a situation in the past when there was a similar value for the known parameter(method-2)? The answer is yes.. There were two days when the value was zero but the precaution was taken on Friday with an higher quantity (1.5). But in the current scenario(position: n), the n-1 and n-2 values aren’t equal or higher than the value 1.5. So the machine understands that the precaution was not taken in this case. Hence the value should meet up the day’s demand.
Hence the machine would calculate the mean(average) value for the Day’s total, considering it to be the average consumption/required quantity for the day. How to calculate the mean?
The mean required quantity per day is 1.6 litres. Hence the machine would settle with the value of 1.6 which closely associated to the value (1.5 litres) that I personally calculated while getting the milk from the milk man.
Hence as per my understanding, Machine Learning is the process of training a system with the data or values from the set of observations from the past and give it the freedom to choose/predict/act based on a best suitable algorithm to handle the situation.
This is my first tech article. I have shared my knowledge with a real life example so that it could be easily understandable and helpful to you all.. If there is any concern or if you can help me out to improvise my understanding and the article, feel free to write to me at seekershiveminds@gmail.com.
