Posts

Unsupervised Learning

Image
Unsupervised Learning: An Unsupervised Learning is where the Machine learns by itself when provided with a data with same kind of labels. The below image is an example of Supervised learning where the Labels (O & X) are clearly specified given 2 Features/Attributes. The distinguishing factor is clearly slated and the Machine is asked to learn it. But in the case of Unsupervised learning, the dataset that is provided doesn't have a distinguished label and the Machine is asked to make some sense of it. The Machine tries to understand the data and makes some sense of it by Clustering them together that lie nearby as below. This is a clear example of Clustering , where closely related labels are grouped together. A very good example of Clustering Problem is Google News. The closely related news are grouped together so that the Users can read all the news that relate to it. Below we can find the news related to BP Oil firm. Q&A by Andrew Ng: Unsuper
Post a Comment
Read more

Supervised Learning

Image
Supervised Learning: We provide a dataset of right answers and ask the Machine to get trained on those datasets, later on once the Machine builds a Model out of those datasets, we ask the Machine to predict the outcome when provided with similar data. In simple terms, the Machine first learns from the given data and then predicts the outcomes based on that. 1) Housing Price Prediction: Andrew Ng used Housing Price Prediction problem for the same which is considered as a Regression Problem. The dataset contains 2 columns, Size in sq. ft. and Price in $1000. Now when the Machine gets trained with this dataset and we can ask the Machine about what would be the Price of the House given the Area of the house. In here also Size in sq. ft. is the ONLY Feature/Attribute. Pink Line denotes a Linear equation Blue Line denotes a Qudratic equation with polynomial 2 Let's not worry about how to differentiate about which one to choose now. 2) Breast Cancer Prediction: In t
Post a Comment
Read more

What is Machine Learning ?

Image
Old Machine Learning Definition: Field of study that gives computers the ability to learn without being explicitly programmed. By Arthur Samuel (1959) The above statement by Arthur Samuel is considered an Old one. A new one was stated by Tom Mitchell, a friend of Carnegie Melon. New Machine Learning Definition:  A computer program is said to learn from experience E with respect to some task T and some performance measure P , if its performance on T, as measured by P, improves with experience E. By Tom Mitchell (1998) Experience E: I think this is the data collected overtime. Task T: Action from the experience Performance P: Probability of right action A Question asked by Andre Ng in his class. Types of Learning: 1) Supervised Learning:  We teach the computer how to do something 2) Unsupervised Learning: The computer learns by itself.
Post a Comment
Read more

Introduction to Machine Learning

This blog is purely for Study purpose, it is created by taking notes from Andrew Ng's Coursera Course.
Post a Comment
Read more