关于rasa_nlu 的regex_features
我想请问一下他是如何工作的，他是怎么样设置nlu和pipline并且如果匹配错误会返回什么。我这边的设置应该做怎么样的一个修改，万分感谢！
这是我设置的内容
{
"rasa_nlu_data": {
"common_examples": [
{
"intent": "zipcode",
"entities": [
{
"entity": "zipcode",
"value": "52231",
"start": 6,
"end": 11
}
],
"text": "这是我的邮编52231"
},
{
"intent": "zipcode",
"entities": [
{
"entity": "zipcode",
"value": "34212",
"start": 0,
"end": 5
}
],
"text": "34212是我的邮编"
}
],
"regex_features": [
{
"name": "zipcode",
"pattern": "[0-9]{5}"
}
],
"lookup_tables": [],
"entity_synonyms": []
}
}pipeline:
- name: "MitieNLP"
model: "data/total_word_feature_extractor_zh.dat"
- name: "JiebaTokenizer"
- name: "MitieEntityExtractor"
- name: "EntitySynonymMapper"
- name: "RegexFeaturizer"
- name: "MitieFeaturizer"
- name: "SklearnIntentClassifier"
- name: "CRFEntityExtractor"但是跑出来的结果不是我想要的。不论我输入正确的邮编还是错误的邮编我都没有看到任何regex工作的信息
Next message:
3422是我的邮编
{
"intent": {
"name": "zipcode",
"confidence": 0.37416135730113564
},
"entities": [
{
"entity": "zipcode",
"value": "3422",
"start": 0,
"end": 4,
"confidence": null,
"extractor": "MitieEntityExtractor"
}
]
"text": "3422是我的邮编"
}