Deciphering Machine Learning
Machine learning is, probably, the buzzword in almost all the industry. Be it IT, pharmaceutical, retail, you name it. People have started boarding the ship of machine learning without prior knowledge. I feel this is good and bad. But, with this piece of knowledge, I will throw some light on machine learning, especially for newbies.
Welcome to the world of machine learning. It's easy to see why the word is making a lot of noise around it. Let me express myself about my views on machine learning in the simplest of the literature. Machine learning is a field of computer science that gives computer systems the ability to "learn" (i.e., progressively improve performance on a specific task) with data, without being explicitly programmed (Wikipedia)
I divide the entire machine learning into two broad categories 1) Supervised ML (Machine Learning) 2) and Unsupervised ML. There are other pre-processes which are involved in data mining, data manipulation etc. In the categorization, I have classified the algorithms.
Supervised Learning: In this type of the algorithm, we supervise (learn) the previous records of the data to understand the demographic, behavioural and/or transactional features of the individuals/users. This data is then used to predict the values in the future, assuming all the conditions are the same.
Don't we do this, in our real life? If we friends meet up, and we have that one friend who always says no to a movie, we might not ask him again. Using his previous behaviour, we predicted his answer and didn't bother to ask him. Organizations across the globe use these algorithms to identify potential customers and to retain existing ones.
Unsupervised Learning: In this type of algorithm, we do not predict the values, but basically, try to identify similar patterns of different individuals and club them into one. Clustering is the easiest example of this.
We know that two of our friends in the group like soccer (Football), three like cricket and the other two like snooker. We have categorized our friends into these groups using their love for the sport. Using these models, companies float their offer to their prospective customers, an important part of marketing analytics.
Majority of the algorithms can be classified into these broad categories, I would suggest all the newbies learn and understand the science that is involved. This will make the game even more interesting.