Just wanna share my experience

Mixture distribution mixes my mind….

Cocktail I thought that working on many sophisticated formula already finished last trimester….
But I was totally wrong…it getting worse….
This trimester I have to face Advance Modeling techniques (AMT). Just like its name, in this course we are dealing with special, advance problem which not come in generic statistics situation.in simple statistics, we usually want to know weather two population are the same, or if we know that they are different we wish to know how to distinct/differentiate them.
If the situation is on the way around, we have a set of data, and we want to know from how many population is the set of data, and from which distribution they are..you must know AMT.
The basic of this course is mixture distribution. In simple statistics we assume that data comes a single normal distribution. So we can make inference from the statistics (value that we have from dataset, eg. mean sample). Practically, the distribution of the data is not always normal, even we can use Central Limit theorem
to approximate the distribution using normal distribution.
But if the case that the distribution have two modes (bimodal) it seems that the distribution is a mixture of two distribution, eg. mix of two normal dists and jointly make that shape…

I will explain in common words, We are served a glass of cocktail and want to know the ingredients of it, how much sugar, how many percentage of milk in it and more difficult is how many ingredients on it, what are they and their percentage..
In our case we deal with data. How many distribution are involve on the data set , what are they and how much their participation….

Confusing right? actually the general idea is not confusing, the problem for me is the mathematical background behind it is a major problem..
wish me luck friends to cope this course……


May 11, 2008 Posted by | Statistics | | Leave a comment