Data mining consists of logical processes to search through large volume of data in order to find important information. At the end this technique will help to find certain patterns that were previously unknown. Once you have found these patterns, you can use them to solve a number of problems. Once a particular pattern is observed, it means that we could be able to predict the behavior of something similar to what we are analyzing.

This will help to make strategic decisions for an organization that can allow you to achieve certain goals. To achieve the end results as mentioned above, the whole process follows certain distinct stages. They are:

• Exploration
• Pattern Identification
• Deployment

Data mining Stages

Stage-I: Data Exploration

In this stage we try to recognise out important variables that are relevent to the problem we are addressing and the nature of these variables. Typically this is how this stage of data mining is realized:

Clean the data

It is often required to clean the raw data we have. Usually the data is transformed to another form to make it easy for rest of the stages.

Create subset of data

It is not good to keep whole of the data we got in the beginning and is more convinient to deal with subset of this data. Instead of considering large number of variables we need to reduce them to a range that is easy to deal with.

Making predictions

We need to make high level analysis of data. The objective here is to make some predections depending on the problem we are dealing with. For doing this we can use certain tools like graphs and statistics.

Stage-II: Pattern Identification

This stage of data mining can be somewhat complex. Objective here is to look for patterns in the data we are analyzing and choose the one which will help to make best prediction. This stage of data mining can be done in different ways, best among them is to apply all different patterns to given situation and choose the one which performs at the top level.

Here is an example for this process in real world scenario. Suppose a super market store wants to make more profits and pattern identification can be best achieved by making 2-3 different shopping patterns for customers and then apply them to hypothetical stragety to determine which one performs best.

Stage-III: Deployment

Once we found a consistent pattern from stage 2 that is highly predictive, we will go for deploying it. For example, if many of the customers are consistently buying a specific product on a certain date, we will be able to predict their future behavior. Now that you've done this, you can take the pattern and apply it in order to see if you can achieve the desired outcome.

At the end of these stage what we achieved is the INTUITION. We now have these institutions with knowledge that will allow us to make strategic decisions in a situation that is not certain.