Skip to content

Rule-based Topics

The new rule-based topics allow you to create a topic that meets certain conditions. It provides higher granularity and can help you understand your data better. It consists of two sections, meet ALL of the following conditions and meet ANY of the following conditions.

Conditions in meet ALL of the following conditions use the “AND” operator. This means that each condition in this section must be true for a topic to return useful insights. Conditions in meet ANY of the following conditions utilise the “OR” operator. This means that at least one condition in this section must be true.

The ALL and the ANY condition sets can be used in conjunction. When using both, a topic must match every condition in the ALL section, plus at least one of the conditions in the ANY section. If there is only a single condition in the ANY section, it must be true.

Conditions

Each condition consists of three fields: operator, type and term/topic.

Operator

There are two types: sentence and text. Sentence operates on utterances, it goes through one sentence at a time and checks if the condition is met. Text, on the other hand, operates on the entire text.

Type

The available types are Topic Negated, Topic Contains, Topic Not Contains, Term Equals and Term Not Equals.

Term

Select an existing term or create a new one when the type selected is Term Equals or Term Not Equals.

Topic

Select a term-based topic from the existing topic list when the type is Topic Negated, Topic Contains or Topic Not Contains.

Examples

In the image below, we created a rule-based topic that will include the topics "hotel" and "location". Conditions including topics hotel and location

When we have a look at the examples, we see that both topics are present in the texts. Examples of rule-based topic that includes topics hotel and location

In this case, we want to find those texts where the term "staff" appears negated by using the type Topic Negated. Examples of rule-based topic that includes topics hotel and location

Here we can see examples where "staff" is negated: "there are no staff there!", "lacking by front-desk staff" and "there weren't enough staff". Examples of rule-based topic that includes topics hotel and location