A guide to building a kickass knowledge base
Mankind has made progress in leaps and bounds in his manner of living, and of course in the technology he uses. So we have gone through the Copper age, the Bronze Age and now have reached the information age. This is the age of the information superhighway. Gen X thrives all things based on knowledge-based systems. Data, or more specifically knowledge reigns supreme in all walks of like. With Artificial Intelligence making an appearance on the scene, the importance of properly accumulated and represented knowledge is at an all-time high.
What is a knowledge base
In any Knowledge-Based System, knowledge is created or acquired from various sources and maintained separated from the inference engine. The Inference engine is a module responsible for processing the data available and deriving conclusions from the knowledge base. The knowledge base contains facts, rules and relationship between data objects in a specialized domain.
Because the inference machine is a wholly separate entity, the same inference engine can be used to analyse different knowledge bases and come up with totally unique conclusions. This increases the strength of the inference engine and makes it multi-functional.
The architecture of a Knowledge-Based System typically includes an I/O interface, an editor, an inference engine, a memory element and of course the knowledge base. Some more elaborate Knowledge-based systems include explanation and learning modules and case history files.
How to create a knowledge base
The process of creation of a knowledge base involves basic steps, which can be group into two major ones, namely knowledge acquisition and knowledge representation.
Here’s how a knowledge base can be built:
- Decide on the main domain and identify the core elements of your knowledge base.
- Decide which topics to start with
- Plan out the structure of the content.
- Assimilate and put data together
- Add diagrams etc if and where required
- Analyze the knowledge base and improve upon it
Once the main domain is identified the knowledge engineer has to implement the knowledge base in a computer-based system. In order to perform this task, relevant data points and knowledge have to be procured from an expert in that particular topic. Data also needs to be acquired from any other relevant sources, including books, papers, reports and so on. Being that we are now riding the information superhighway with all the resources available online, this should be no daunting task.
Although knowledge acquisition is the most time-consuming and costly part of the building process, the knowledge engineer can take advantage of the editor module of the KBS. Using this tool, the KE can save a lot of time since it provides mechanisms for validation and verification, completeness and consistency.
How to choose knowledge base topics
The topics in the knowledge base must be the relevant ones that are essential to the analysis and presentation of reports. For example in an analysis of a disease, there are many parameters of the human body that need to be included in the knowledge base to be accurate. In such a case single topic missed could literally be the difference between life and death.
Structuring the knowledge base
In order for a knowledge base to be usable by an inference engine, its structure must be suitable for its use. The structure is key to having a killer knowledge base. A solid Information Architecture is the key that tells you how to structure the knowledge base, and how each piece of data should fit into that structure. Not only will it help work efficiency, but the content will be stronger as a result.
Knowledge base templates and examples
The created knowledge base is probably full of beautifully written content, but unless there’s an efficient Information Architecture in place, that content won’t ever live up to its potential.
Knowledge base tools and software
Here are a few knowledge base software available in the market.
Of course these are just a few of the many software available. The choice of software is dependent on factors like utilization, the purpose of use, and of course, cost.
Mistakes to avoid when building a knowledge
- Inaccurate data/knowledge: For example, if the knowledge base is in a specialized area of medicine, the relevant information and data need to be sourced from experts in the field as also from relevant medical journals. Because the knowledge engineer should almost an expert himself in the chosen domain, the acquisition of knowledge is a very demanding task. Also, for the knowledge to be usable, it must be accurate to a high degree and presented in the correct format to the inference machine.
- Incomplete data: The knowledge must also be complete in the sense that all essential facts and rules are included and free of duplication.
- Mismanaged mobile services: Given that most end users access services via mobile platforms, its vital that the mobile platform should be sufficiently intuitive and responsive to the needs of the end customer.
- Too big, too fast: it is important that the knowledge in a knowledge base is structured in stages for ease of access by the user.