Identify sentences with multiple information.

Customers have a habit of providing a lot of information in one sentence. The problem is that the Bot needs to accurately extract the information in a sample sentence and correctly identify its entity

Composite entity

For information appearing in sentences, belonging to different types of entities, the trainer uses the Composite Entity training feature according to the following instructions:

• Step 1: Select the Entity Type tab (1) then click Create New (2) to add a new entity

• Step 2: At the display interface, select Composite (3) in the Entity Type section.

• Step 3: Name the composite entity type (4) and click Add_Pattern (5) to add a cluster of entities that need to be identified. The structure consists of many child entities. Click save to complete (6).

The system will check all pattern values ​​in order of priority from top to bottom.

For example
	Consider the sample sentence: "I want to order 5 cups of milk tea size M" The system will check with pattern options in order of priority from top to bottom of the list. Accordingly, the system will extract information in the following order: #sys.number #drink #drink_size In this example: 5: sys.number milk tea: drink size M: drink_size

After creating the Composite Entity entity type, perform training on the sample sentence as follows:

• Step 1: At the Sample Sentences Tab, select Create a new sample sentence.

• Step 2: At the displayed interface, select Intent (1) if you want to create a sample sentence, enter the sample sentence in the content box (2). Highlight the information cluster that needs to be trained for the composite entity type (3).

• Step 3: Click Save (4) to complete adding the sample sentence.

After each step of adding new data, click the Training icon at the quick access toolbar to train new knowledge for the Bot.

Parameter setting (List)

Setting list parameters allows the system to recognize sentences with multiple information belonging to the same entity type.

For example:

Consider the sample sentence "I want to order 5 cups of milk tea and 4 cups of peach tea" with the information that milk tea and peach tea both belong to the entity type "drink". The trainer sets the parameters as follows:

Sample sentence	Intent	Entity	Parameters
“I want to order 5 cups of milk tea, 4 cups of peach tea”	Order	milk tea (drink) peach tea (drink)	drink

-> Setting the "List" option for the “drink” parameter helps the Bot identify and extract information about the drink entity type in the sample sentence