Description of detection operator objects
This document describes the structure of detection operator objects in rules. These operators include checks for text, speech, emotion models, and call parameters.
Description of the operators parameter
HIT_ANY_KEYWORDS: Contains any keyword
INCLUDE_KEYWORDS: Contains all keywords
INCLUDE_KEYWORDS: Contains any N keywords
INCLUDE_KEYWORDS: Contains none of the keywords
Property | Value type | Example value | Description |
contextChatMatch | boolean | false | Specifies whether to analyze multiple sentences. |
in_sentence | boolean | false | Specifies whether the rule applies within a single sentence. |
keywordMatchSize | int | 1 | Contains N keywords |
keywords | String array | ["Hello","hello"] | An array of keywords. |
threshold | int | 4 | The total number of keyword hits. |
SIMILAR_MATCH: Similarity match
Property | Value type | Example value | Description |
score | int | 80 | The similarity score. |
similarlySentences | String array | ["hello","world"] | Similar phrases. |
in_sentence | boolean | false | Specifies whether the rule applies within a single sentence. |
REGULAR_EXPRESSION: Regular expression check
Property | Value type | Example value | Description |
regex | string | [a-z0-9_-]{3,16} | The matching expression. |
notRegex | string | 1[3|4|5|7|8][0-9]{9} | The exclusion expression. |
in_sentence | boolean | false | Specifies whether the rule applies within a single sentence. |
ADVANCED_REPEAT_DETECT: Contextual repetition check
Property | Value type | Example value | Description |
from | int | 3 | Repetitive content within N sentences |
similarity_threshold | float | 0.6 | Similarity |
hit_time | int | 3 | Nth iteration |
threshold | int | 5 | Content that contains fewer than N characters is not detected. |
excludes | String array | ["hello","world"] | Exception Dialog Box |
NER_MODEL: Information entity check
Property | Value type | Example value | Description |
customerParam | Object | Object | The entity information. |
dataType | int | 1 | The entity type. |
INTERVAL_GREATER: Call silence check
Property | Value type | Example value | Description |
interval | int | 2000 | The silence duration in milliseconds. |
different_role | boolean | true | Specifies whether to distinguish between roles. Set to 1 to distinguish roles, or 0 to not distinguish. |
target | int | 1 | Specifies whether to detect the same role. Set to 1 for the same role, or another value for different roles. |
from_end | boolean | false | |
compareOperator | string | gt | The comparison operator. `gt` means greater than. `lt` means less than. |
SPEECH_SPEED_CHECK: Speech speed check
Property | Value type | Example value | Description |
velocityInMint | int | 2000 | The speech speed in words per minute. |
minWordSize | boolean | true | Content that contains fewer than N characters is not checked. |
average | boolean | false | Specifies whether to check the average speech speed for the entire conversation. |
GRAB_WORDS: Interruption check
Property | Value type | Example value | Description |
interval | int | 2000 | The interruption time in milliseconds. |
threshold | int | 4 | The number of words in the interruption. |
delayTime | int | 2000 | The time in milliseconds to wait before determining an interruption. |
ROLE_CHECK: Role check
Property | Value type | Example value | Description |
target_role | string | Customer | Customer and customer service |
DURATION: Abnormal call handling
Property | Value type | Example value | Description |
compareOperator | string | gt | The comparison operator. `gt` means greater than. `lt` means less than. |
interval | int | 5000 | The hang-up time in milliseconds. |
beginType | string | DIALOGUE | Recording and Dialogue |
endType | string | DIALOGUE | RECORDING DIALOGUE |
target_role | string | Customer Service | The role to detect. |
DURATION: Recording duration check
Property | Value type | Example value | Description |
interval | int | 2000 | The recording duration in milliseconds. |
compareOperator | string | gt | Greater than (gt) and less than (lt) |
ASR_EMOTION: Energy detection
Property | Value type | Example value | Description |
maxEmotionChangeValue | int | 5 | The energy value. |
checkType | int | 3 | The check type. 1: Energy fluctuation between adjacent sentences. 2: Maximum energy span. 3: Energy range detection. |
compareOperator | string | gt | The comparison operator. `gt` means greater than. `lt` means less than. |
DIALOGUE_SIZE_CHECK: Maximum dialogue sentence count check
Property | Value type | Example value | Description |
interval | int | 5 | The number of sentences. |
compareOperator | string | gt | Greater than (gt), less than (lt) |
CUSTOMER_CHECK_MODEL: Customer model check
Property | Value type | Example value | Description |
pvalues | String array | ["tousu01","zhaozhuguan01"] | The detection type. |
Customer model enumeration values
tousu01 | Complaint |
zhaozhuguan01 | Escalate issue |
chaping01 | Threaten to give a bad review |
zhiyitaidu01 | Question the service |
customer_praise01 | Praise |
dafei01 | Question the agent for giving an irrelevant answer |
internal_complain01 | Internal complaint |
public_opinion | Create public sentiment issue |
ruma01 | Abuse |
fennu01 | Anger |
ABUSE_MODEL: Agent model check
Property | Value type | Example value | Description |
pvalues | String array | ["tousu01","zhaozhuguan01"] | The detection type. |
Agent model enumeration values
fanwen01 | Follow-up question or retort |
jifeng01 | Follow-up question or retort (sarcastic) |
service_guide_complain01 | Guide to complain |
service_buckpassin | Shirking responsibility |
didnot_pacify | Failed to pacify the customer |
ruma02 | Abuse |
TYXM_LLM: Custom detection conditions
Property | Value type | Example value | Description |
type | string | TYXM_LLM | A fixed value of "TYXM_LLM", which indicates the custom detection condition operator. |
sceneName | string | Customer Service Scenario | The scenario name. |
knowledges | Knowledge array | Background knowledge (optional). | |
dimensions | Dimension array | The detection dimensions. | |
hitCondition | int | -1 | The logical relationship between dimensions. -1: Meets any dimension. -2: Meets all of the above dimensions. |
Description of the Knowledge object
Property | Value type | Example value | Description |
dimension | string | Knowledge Point A | The name of the knowledge point. |
desc | string | Specific description of Knowledge Point A | The description of the knowledge point. |
Description of the Dimension detection object
Property | Value type | Example value | Description |
dimension | string | Shows impatient behavior | The dimension name. |
desc | string | Whether the agent shows impatient behavior | The dimension description. |
includeCondition | string | The agent impolitely interrupts the customer.... | The hit condition. |
excludeCondition | string | The exclusion condition (optional). |