Pattern recognition focuses itself in finding trends, patterns and regularities.
Profiling refers to a state, resulted from a construction of data from a data store.
In the application of retrieving information (pattern recognition) from data stores in order to relate nodes (profiles) with each other.
In this article I’m not referring to machine learning or any other academical, mathematical approaches.
In this discussion, a profile might be compared with a state, or a stateful object. Such profile has properties, which contain data, like name, date of birth, location, etc. and together it forms information for the operator.
Furthermore, the profile or stateful object is related to other profiles by means of direct, fixed and static links or by pattern recognition; the attempt to dynamically link the profile to another profile.
The profile is generated from unrelated, not categorized data (records), stored in a data store.
For an organization like NSA, the data store refers to the collection of data collected from everywhere and anyone in order to detect threats or possibilities.
For a commercial organization, the data store refers to the network and its internal processes in order to detect worms, virus software, suspicious software and employees.
For a scientific organization or institution, the data store refers to a collection of observations in order to improve and increase scientific knowledge and achievement.
A system like this needs a few people and a generic system to retrieve information.
But what happens in organizations like NSA, GCHQ, FSB, 3PLA and others?
They have hundreds, even thousands of operators attempting to tag, categorize and prioritize all that huge amount of data, hoping to increase speed of retrieval.
They use standard, commercial software products to attempt to retrieve that data with the help if the tags, categories and priorities.
They work with their stomach instead of their head. They ‘dirty the fatal stores with a potential treasure of potential information.