I am in deep need of a framework for calculating joint and conditional probability tables from a simple array of multivariate data. I need these for computing in probabilistic graphical models. Can anybody help me out?
I think you must consider the following questions: Is your data of continuous or discrete nature? Is your data following a parametric or a non-parametric distribution? Would you like to estimate the PDF?
The easiest way to calculate an estimated probability from a data set is to divide the data into classes. I would recommend you to start analysing the dataset using the Distribution fitting tool, so you can select the right bin size for you histograms. Then, you can build the histogram of relative frequency and estimate the probability from it. You can use the hist() function or the hist3() function for bivariate data. You can also search in mathworks for customized functions for a generalized multivariate analysis. Then the join probabilities could be easily calculated. Then, using the bayes theorem you can easily obtain the conditional probabilities.
Hi, Thanks for your thoughtful suggestions. Actually I am asking about what is the way of implementing these probability calculations in computing s/w s like MATLAB...What should be the data structure to use in these computations to make the program not computationally challenging. Thanks again.