List lemmatized classification data
This endpoint lists the lemmatized classification data used for training the chatbot. It consists of lemmatized chat phrases that a user could potentially type, along with their associated reply category.
Note the following two advantages of lemmatizing the training data:
-
The input message data is also lemmatized and therefore aligns more closely when calculating a match.
-
Less training data is required because fewer variants of words are required: variants converge onto their lemma.
Response
|
Data is truncated for readability. |
Notice that the data is lemmatized:
Response
{
"lemmatizedData": [
{
"category": "animal-list",
"lemmatizedPhrases": [
"what animal do you have"
]
},
{
"category": "directions",
"lemmatizedPhrases": [
"where be you"
]
},
{
"category": "opening-times",
"lemmatizedPhrases": [
"when be you open"
]
}
]
}