I am currently studying the Predictive Analytics. There are two questions which I couldn't find the answer is:
(1) For what predictive analytics (such as Neural Network, Case Based Reasoning etc.) we need to normalize the data, and what is the purpose of that?
(2) Which type of data is suitable for which predictive analytics? (For example, how is the type is suitable for using Neural network, or Case Based Reasoning?)
Please give me some ideas or recommend some good sources I can read? I couldn't find articles which compare the uses of analytic models
Thank you
Dear friend,
A helpful link:
https://en.wikipedia.org/wiki/Predictive_analytics
this might be a short/idiot-proof reading... to refresh key ideas...
http://www.cs.ucf.edu/courses/cap6412/fall2009/papers/Berwick2003.pdf
You are most welcome, although I am no Doctor nor Esq. Some website professional association bots love to 'impose' undeserved title on my profile... I have since given up on 'correcting' automated tools with my feeble human means.
Well, the main difference between CBR and NNs or SVMs is that in CBR you do not learn a model. It is what is known as a "lazy learner". In classification tasks is like a Nearest Neighbor classifier. For further info you may look at this paper:
R. López de Mántaras, E. Armengol; Machine Learning from Examples: Inductive and Lazy Methods. Data and Knowledge Engineering Journal, 1998, Vol. 25, nº 1-2, pp. 99-123.
- Badges
- Science topic