Structured and unstructured data (2)

For communication, thinking, and reasoning abstractions and structures are important. Humans started using both structuring and abstractions. It appears that most advances in human society are somehow related to the ability of creating new abstractions and developing new structuring processes. Structuring processes are related to abstractions, but they are fundamentally different and should be treated so.

While trying to understand a new domain or application in sciences of any kind – whether physical or social sciences – one initially faces completely unstructured data. There are no concepts or measurements that will allow thinking and analyzing the data related to the application. To understand the domain one starts looking for familiar concepts from a familiar domain that may have some similarity with this new domain. If found, then one adapts those concepts, otherwise needs to invent new concepts.

In most cases, one can consider abstractions and structuring as two different approaches used to understand, represent, and share knowledge and experiences. Without abstraction and structuring, it is rarely possible to share and represent knowledge and experiences.

Abstractions and structures play different yet often complementary roles. Abstraction allows one to identify properties or characteristics that are shared by several objects, concepts, or events that could be considered belonging to one class. This is the process of going from concrete to abstract. Out of several characteristics of an object, one must identify those that are essential for the membership in this class. Obviously, an object sometimes could belong into multiple categories depending on the characteristics it has. The abstraction is the process of identifying ‘essential’ characteristics that an object must have to belong into that class. This is not an easy task.

Structuring is the process of taking a particular attribute space and imposing some structure on this. The structure imposed may have some meaning but is really imposed by humans to easily talk about the attribute space as well as provide some kind of order on that space. Take time for example. Time is a physical attribute. To be able to talk about time different calendar systems were developed in different parts of the world. These systems were inspired by some physical observations but after a level were quite arbitrary – such as the concept of hours, minutes, and seconds. One could make similar observations about other attributes also.

Leave a Reply

Your email address will not be published. Required fields are marked *