There are many challenges in knowledge classification. In biology, every organism falls under a certain hierarchy of classifications, or taxonomy. However, modern knowledge management work can be more demanding. Suppose for example you want to classify your customers by their industry and their field of work, you may end up with many anomalies and surprises.
For example if you want to classify these companies:
- a real estate brokerage company
- a construction company
- a building materials company
Can you see the overlaps between their business? Can you put them under one category? How about if a company is doing a bit of all of these aspects.
The problem becomes evident when you do the classification with a certain goal in mind, such as finding or grouping suppliers or customers with a specific criterion. Then you’ll find some of the items don’t fit any of the categories. Alternatively, you can find categories which are too narrow, or too broad! What can you do in a situation like this?
The problem with “Other”
First of all, it is important to look into the classification system: that you are aware of all the features and variations in all the items you are classifying. It is important to note that there can be always outliers, which can form the basis of what can be the continuous development of the category system. Imagine that you have a form that gives you a few dropdown options, but none of these apply to you. In this case, you will need to have “other” option. This “other” category is both a challenge and an opportunity.
One of the challenges of “other” category is that it is loosely defined. In addition, we cannot really search something with “other” criteria. However, it can present an opportunity. The opportunity is that when there are more items under “other”, and we collect in a free form text what they are: we can start forming new category structure.
Items falling under multiple categories
This is a duck:
How can you classify it:
- things that fly
- things that swim
In many cases, we want to capture the flying aspect. However, in others we want to capture the swimming aspect. And in many contexts, we are mainly interested in the classification as a domesticated bird, and/or its food value.
Therefore, classification is indeed a difficult problem. No classification can capture all aspects of the item to be classified. The only way is to allow applying multiple classifications to the same item.