Knowledge Bases are required inputs for Insight Booster projects. Once you have created a project, open it and click on the “Link” button to connect a Knowledge Base. To learn how to create a Knowledge Base, please refer to our article: Creating a Knowledge Base
From the drop-down list, select one or more Knowledge Bases to initiate your project. Once you have made your selection, click “Continue”. You can add or remove Knowledge Bases from your Insight Booster project at any time.
Metadata fields in your dataset can be used as filters to refine your analysis. For example, if your Knowledge Base includes country metadata, you can tailor your output by selecting specific countries from the drop-down menu. If you wish to proceed without any filters, simply click “Continue”.
To refine your search, enter a few words describing the topic you are interested in. If all the texts are relevant, simply leave this field empty to use the default setting. After making your selection, click “Apply”. If you want to limit your Knowledge Base to a specific area of interest, begin by describing it. To reiterate, only fill out this section if you want to filter your Knowledge Base.
Topic of interest: Provide a brief description of the subject you are focused on.
Limit the matches to the following number of most relevant texts: Specify the number of most relevant texts to include, with a maximum limit of 10,000. This determines the total number of texts used in the Knowledge Bases.
Only include texts that meet the relevance requirement: Set the criteria for text relevance to determine how much extraneous content (or "noise") you are willing to permit.
You will be back on your Insight Booster project’s Knowledge Base section and will see the Knowledge Bases that have been linked to your project. With these resources connected, you are now ready to leverage the full analytical capabilities of the Insight Booster for your data analysis.
To get an overview of the data visually through a landscape, please refer to our article: Generating a content landscape and using the Research Chatbot