Question-16: Can you give some more detail about supervised learning?

Answer: We can describe supervised learning as task-oriented and which is highly focused on the singular tasks. And you should have as much as possible example data to train your model, so that in production it can predict dependent variable more accurately.

Question-17: Can you give few examples of Supervised Learning?

Answer: Yes, below are the few examples

• Gmail Auto classifications: You have seen in your Gmail box Google classify the emails whether an email is for marketing, social networking, spam or most important etc.
• Spam filter: Your new email based on contents can be classified whether it’s a spam or not. As well also check your response as well to each email. Yes, email provider reads your email to correctly classify it.
• Face recognition: Now in smart phone, you have seen that Camera automatically capture the face. So, your mobile has Supervised Machine Learning trained model in it and while you capture photos, they can suggest faces etc.

Question-18: What is un-supervised learning?

Answer: You can think of its an opposite of supervised learning. It does not have pre-defined labels. (There is nothing which should supervise the Model). But you would feed whatever data you have to Machine Learning Model and based on the features and characteristics it generates the groups or you can say cluster. Which you can easily understand the way data is organized or grouped by Model.

Question-19: Why unsupervised learning is more challenging?

Answer: If you see most of the data currently is unlabeled (there are various jobs opening in recent years to just label the data, so that supervised learning can be used). And it is very interesting to have such an intelligent Machine Learning model which can accurately groups the Tera-bytes of unlabeled data. And make sense out of these unlabeled data is very demanding because these data were lying idle and previously was of no use and was discarded. But now that is not the case, every company want to make intelligent decision based on this idle unlabeled data as well. And day by day new unlabeled data is being generated.

Question-20: Can you give an example of unsupervised learning?

Answer: Yes, lets assume you have 1 Million research papers on various topics. Now you want to group all research papers in respective research areas. However, its not known in advanced what all research areas these papers are written. Hence, you will be running one of the un-supervised Machine Learning algorithm which can group all these papers in respective or related research areas.