Question

Business analytics MBA- There are a number of learning scenarios or types of learning algorithms, that can be used depending on whether a target variable is available and how much labeled data can be used. These approaches include supervised, unsupervised, and semi-supervised learning. Explain the difference between each type of machine learning. Give an example of how each is used. Write your responses in detail with examples. Be sure to identify the source of your example in your posting. Your initial post should be of minimum 300 words.

Answer 1

Machine learning is also know as predictive analysis or predictive modelling which describes the ability of the computer models to learn without explicit programming. The basic programming is done to receive and process the input and predict the output with improvisations being modelled with the analysis done on the input.

The four major machine learning methods are the supervised learning, semi-supervised learning, unsuoervised learning and the reinforcement type of learning. In the supervised model, the operator feeds the model with the necessary input and ouput programming and the computing mechanism would be entrusted with how to arrive at those outputs with the given inputs. Classification, regression and forecasting are the major examples for supervised learning. For example, the classification of the mail contents in to spam and non-spam contents can be considered under this category. Unsupervised learning is similar to the supervised learning strategy with the variance that it also consists of unlabelled inputs along with the labelled inputs and hence the algorithm would be able to mark the label from the unlabelled patterns.

In unsupervised learning, the correlation and analysis is being done by the machine itself as it doesnt have a supervision for the specific output patterns and hence the machine has to find tha data analytics by itself. Clustering and dimension reduction are the major examples under this category. In reinforcement learning, the machine is being provided with a set of actions, parameters and end values. Thus trial and error can be obtained as a part of this learning method to get the reaponse to a situation.