CN-121980011-A - Auxiliary generation system and method for mediation strokes based on artificial intelligence

CN121980011ACN 121980011 ACN121980011 ACN 121980011ACN-121980011-A

Abstract

The invention discloses an artificial intelligence-based auxiliary generation system and method for a regulation record, which relate to the technical field of auxiliary generation of regulation records and comprise the steps of analyzing the correlation degree between regulation indexes and regulation materials in a key index library, analyzing the similarity degree between the history regulation records in different history regulation case sets, evaluating the change situation mutation conditions of target regulation indexes in different time spans in a regulation case database, obtaining characteristic regulation indexes, determining the target time range of the characteristic regulation indexes for searching in the regulation case database, analyzing the reference value generated by the regulation records of the history regulation case in the regulation case database, assisting in the generation of the regulation records of the regulation materials, ensuring that the generated regulation records not only meet the condition of the regulation materials in content, but also solving the problem of time-lapse failure of the regulation materials, and improving the writing efficiency of regulation records by a regulation person.

Inventors

LI LONG
GU HAO
SHAO WEI
WANG YUE

Assignees

江苏新视云科技股份有限公司

Dates

Publication Date: 20260505
Application Date: 20260407

Claims (10)

1. The auxiliary generation method of the mediation pen records based on the artificial intelligence is characterized by comprising the following steps: step S1, acquiring a mediation material and a key index library from a platform, and analyzing the correlation degree between the mediation index in the key index library and the mediation material to obtain a target mediation index; Step S2, acquiring a reconciliation case database, acquiring a historical reconciliation case set containing target reconciliation indexes under different time spans from the reconciliation case database, and analyzing the similarity degree between the historical reconciliation strokes in different historical reconciliation case sets to obtain stroke similarity data; Step S3, according to the similar data of the stroke records, the change situation mutation conditions of the target mediation index in different time spans in the mediation case database are evaluated, the characteristic mediation index is obtained, and the target time range of the characteristic mediation index for searching in the mediation case database is determined; And S4, analyzing the reference value generated by the adjustment record of the adjustment material of the historical adjustment cases in the adjustment case database according to the target time range, and assisting in the generation of the adjustment record of the adjustment material.
2. The method for auxiliary generation of a mediation pen based on artificial intelligence according to claim 1, wherein the step S1 includes: step S11, acquiring a mediation material uploaded by a mediation person in a platform, and preprocessing the mediation material to obtain a mediation text in the mediation material; dividing the reconciliation text into sentences, extracting words in the sentences in the reconciliation text by using a natural language processing technology, and collecting the words to obtain a word set of the sentences; Step S12, acquiring a preset key index library, and acquiring a keyword set corresponding to each adjustment index from the key index library; acquiring word vectors of words in the word set of the sentence by using a word vector model preset by a platform, and calculating a correlation value between the sentence and the mediation index; step S13, obtaining the correlation value between the sentence and each adjustment index in the platform, and judging that the sentence is related to a certain adjustment index when the correlation value between the sentence and the certain adjustment index in the platform is larger than a preset correlation threshold; Step S14, when the correlation values between sentences and each adjustment index are smaller than a preset correlation threshold value, sequencing each adjustment index according to the magnitude of the correlation value, and calculating the absolute value of the difference between the correlation values between adjacent order adjustment indexes and sentences; when the absolute value of the difference between the correlation value between a certain adjacent order adjustment index and sentences is larger than a preset characteristic absolute threshold value, acquiring the order with higher order in the certain adjacent order adjustment index, marking the order as a reference order, acquiring an h-th adjustment index higher than the reference order from each adjustment index, and judging that the sentences are related to the h-th adjustment index; A plurality of mediation indexes related to the existence of each sentence in the mediation material are obtained and recorded as target mediation indexes contained in the mediation material.
3. The method for auxiliary generation of a mediation pen based on artificial intelligence according to claim 2, wherein the step S2 includes: S21, acquiring a target mediation index in a mediation material, extracting a plurality of sentences related to the target mediation index from the mediation material, and collecting the sentences to obtain a sentence set G of the target mediation index in the mediation material; Acquiring a historical reconciliation case set B from a reconciliation case database preset by a platform, extracting sentences related to a target reconciliation index from a certain historical reconciliation case in the historical reconciliation case set B and collecting the sentences to obtain a sentence set G ́ of the target reconciliation index in the certain historical reconciliation case, calculating a text similarity value S G between the sentence set G and the sentence set G ́, judging the sentence set G and the sentence set G ́ through a set text similarity threshold, and judging that the text similarity exists between the certain historical reconciliation case and a reconciliation material on the target reconciliation index when the text similarity value S G is larger than the text similarity threshold; Step S22, setting unit time length, acquiring a plurality of history adjustment cases similar to the text of the adjustment material on the target adjustment index from the history adjustment case set B, and dividing the plurality of history adjustment cases according to the time length of the distance between the plurality of history adjustment cases and the current period to acquire a history adjustment case set in each unit time length before the current period; Step S23, acquiring a history adjustment case set zeta and a history adjustment case set eta in two unit time lengths adjacent to each other before the current period, acquiring the average value of the feature vectors of all sentences of a history adjustment list in the history adjustment cases in the history adjustment case set zeta, and taking the average value as the list feature vector of the history adjustment list; Acquiring a history feature vector of a history regulation record in each history regulation case in a history regulation case set zeta and a history feature vector of the history regulation record in each history regulation case in the history regulation case set eta, and calculating the average value of cosine similarity between the history feature vector of the history regulation record in each history regulation case in the history regulation case set zeta and the history feature vector to obtain a record text similarity value between the history regulation case set zeta and the history regulation case set eta, wherein the order of the unit duration before the current period of the history regulation case set zeta and the current period of the history regulation case set eta is higher than that of the history regulation case set eta; And obtaining and collecting the similarity values of the stroke records text among the history adjustment case sets in each unit time before the current period of the target adjustment index in the adjustment material, and obtaining the stroke record similarity data.
4. The method for auxiliary generation of an artificial intelligence based mediation pen record as defined in claim 3, wherein said step S3 includes: step S31, obtaining similar data of the strokes of the target mediation index in the mediation material, and calculating a stroke text difference value P of the target mediation index in a unit duration of a certain historical mediation case set; Step S32, when the difference value of the text of the stroke records in any unit time before the current period of the target adjustment index is larger than a preset threshold value of the difference value of the text of the stroke records, judging that the target adjustment index is mutated along with different change situations of the time span; Otherwise, obtaining the maximum value P max and the minimum value P min of the difference value of the stroke text in each unit time before the current period of the target mediation index, and calculating the text difference span L=P max -P min of the target mediation index in each unit time before the current period; Step S33, setting a text difference span threshold L ́, and calculating the variation value of the text of the stroke record of the target adjustment index in each unit time before the current period when L > L ́; When the change value of the stroke text in any unit time before the current period of the target mediation index is larger than a preset difference change threshold value, judging that the target mediation index is mutated along with different change situations of the time span; when L is less than or equal to L ́, judging that the target mediation index cannot be mutated along with different change situations of the time span; acquiring target mediation indexes which can mutate along with different change situations of time spans from each target mediation index of the mediation material, and marking the target mediation indexes as characteristic mediation indexes; step S34, collecting the difference change values of the characteristic adjustment indexes in each unit time before the current period to obtain a difference change set K; constructing each difference range in the difference change set K by taking a minimum value K min in the difference change set K as a starting point and each element in the difference change set K as an end point, acquiring the total number M of the elements in the difference change set K in any one difference range from each difference range, acquiring the total number M sum of the elements in the difference change set K, and calculating the coverage ratio V=M/M sum of any one difference range to the difference change set K; Setting a coverage ratio threshold V ́, when V > V ́, extracting a plurality of difference ranges with coverage ratio larger than the coverage ratio threshold from each difference range, acquiring a minimum value T min of a distance duration between a time period where a unit duration represented by an end point in the plurality of difference ranges is located and a current period, and constructing a time range for searching cases in a historical mediation case set B in the current period by using a characteristic mediation index in the modulating material by using a period between the minimum value T min and the current period to obtain a target time range of the characteristic mediation index in the modulating material in the current period.
5. The method for auxiliary generation of an artificial intelligence based mediation pen record as defined in claim 4, wherein said step S4 includes: Step S41, obtaining a union of target time ranges of all marked target mediation indexes in the mediation material according to the target time ranges of all marked target mediation indexes in the mediation material, and marking the union as a retrieval time range of the mediation material; step S42, obtaining the f historical reconciliation cases of the reconciliation case database in the retrieval time range, and calculating text similarity values between the reconciliation material and the f historical reconciliation cases on the target reconciliation indexes; When the text similarity values of the adjustment material and the f-th historical adjustment case on the target adjustment indexes are larger than a preset target similarity threshold, judging that the historical adjustment record of the f-th historical adjustment case has a reference value for generating the adjustment record, and marking the f-th historical adjustment case as a reference historical adjustment case; And acquiring each reference history adjustment case of the adjustment material, and using a preset intelligent auxiliary generation model to carry out auxiliary generation on the adjustment record of the adjustment material according to each reference history adjustment case and the adjustment material.
6. An artificial intelligence based mediation record assistance generation system for executing the artificial intelligence based mediation record assistance generation method of any one of claims 1 to 5, wherein the system includes a correlation degree analysis module, a similarity degree analysis module, a time range determination module, and an assistance generation module; The correlation degree analysis module is used for analyzing the correlation degree between the adjustment indexes and the adjustment materials in the key index library to obtain target adjustment indexes; The similarity degree analysis module is used for analyzing the similarity degree among the history adjustment strokes in different history adjustment cases of the target adjustment index to obtain stroke similarity data; The time range determining module is used for evaluating the change situation mutation conditions of the target mediation index in different time spans in the mediation case database, obtaining the characteristic mediation index, and determining the time range of the characteristic mediation index in the mediation case database to obtain the target time range; the auxiliary generation module is used for analyzing the reference value generated by the adjustment record of the adjustment material according to the historical adjustment cases in the adjustment case database according to the target time range, and carrying out auxiliary generation on the adjustment record of the adjustment material.
7. An artificial intelligence based reconciliation transcription assistance generation system as defined in claim 6 wherein the relevance analysis module comprises a material processing unit and a relevance analysis unit; The material processing unit is used for preprocessing the mediation material to obtain a mediation text in the mediation material, extracting words in sentences in the mediation text by using a natural language processing technology and collecting the words to obtain a word set of the sentences; The correlation degree analysis unit is used for acquiring a key index library from the platform, acquiring a keyword set corresponding to each adjustment index from the key index library, and analyzing the correlation degree between the adjustment index and the adjustment material in the key index library to obtain a target adjustment index.
8. The artificial intelligence based mediation record assistance generation system of claim 6, wherein the similarity analysis module includes a case set acquisition unit and a similarity analysis unit; The case set obtaining unit is used for obtaining a reconciliation case database from the platform, and obtaining a historical reconciliation case set containing target reconciliation indexes in the reconciliation material under different time spans from the reconciliation case database; the similarity degree analysis unit is used for analyzing the similarity degree among the history adjustment strokes in different history adjustment case sets to obtain the stroke similarity data.
9. The artificial intelligence based mediation transcription assistance generation system of claim 6, wherein the time range determination module includes a mutation analysis unit and a time range determination unit; The mutation analysis unit is used for evaluating the mutation conditions of the change situation of the target mediation index in different time spans in the mediation case database and obtaining a characteristic mediation index; The time range determining unit is used for determining the time range of searching the feature mediation index in the mediation case database to obtain a target time range.
10. An artificial intelligence based mediation transcription assistance generation system according to claim 6, wherein the assistance generation module comprises a reference value analysis unit and an assistance generation unit; the reference value analysis unit is used for analyzing the reference value generated by the adjustment stroke record of the historical adjustment cases in the adjustment case database according to the target time range to obtain a reference historical adjustment case; the auxiliary generation unit is used for acquiring each reference history adjustment case of the adjustment material, using a preset intelligent auxiliary generation model, and carrying out auxiliary generation on the adjustment record of the adjustment material according to each reference history adjustment case and the adjustment material.

Description

Auxiliary generation system and method for mediation strokes based on artificial intelligence Technical Field The invention relates to the technical field of auxiliary generation of a mediation pen record, in particular to an auxiliary generation system and method of a mediation pen record based on artificial intelligence. Background The reconciliation of the record is to record the legal document of the whole reconciliation process in the dispute truly by the reconciliation staff, the traditional reconciliation of record generation is completed depending on the manual record of the reconciliation staff in the reconciliation process and the post-processing of the written form, the reconciliation of the record can be directly generated in this way, but the problem is that firstly, the reconciliation staff is required to reconcile the dispute in the reconciliation process in the actual process, and is required to spend a great deal of time to write the reconciliation record document after the reconciliation is finished, the manpower resources of the profession are wasted easily, so that the quantity of cases to be mediated is limited, more people with needs cannot be served, and secondly, because the traditional mediation writing is manually and purely written by mediation staff, the traditional mediation writing method can lead the mediation staff to have a certain writing level as well as mediation, and the cultivation of such mediation staff needs a long time and cost, and the traditional manual purely writing method definitely causes the waste of a large quantity of manpower resources. Along with the continuous development of technology, it is becoming more and more common to use artificial intelligence technology to assist in generating a reconciliation record, and conventional use of artificial intelligence technology to assist in generating a reconciliation record is mainly to generate a reconciliation record according to a preset database, but in the actual process, the disputed cases of two parties are different, policies and processing methods related to different cases are different, and the reconciliation records are influenced by time, i.e. a case which seems to be a reference may be invalid due to the time effect, so that the reconciliation record does not have a reference value, and a reasonable and missed reconciliation record is easily generated, which not only affects the unable to assist reconciliation personnel, but also may mislead the reconciliation personnel, affect reconciliation progress and even provoke contradiction. Disclosure of Invention The invention aims to provide an artificial intelligence-based auxiliary generation system and method for regulating a stroke record, which are used for solving the problems in the prior art. In order to achieve the purpose, the invention provides the following technical scheme that the method for assisting in generating the regulation stroke records based on artificial intelligence comprises the following steps: step S1, acquiring a mediation material and a key index library from a platform, and analyzing the correlation degree between the mediation index in the key index library and the mediation material to obtain a target mediation index; Step S2, acquiring a reconciliation case database, acquiring a historical reconciliation case set containing target reconciliation indexes under different time spans from the reconciliation case database, and analyzing the similarity degree between the historical reconciliation strokes in different historical reconciliation case sets to obtain stroke similarity data; Step S3, according to the similar data of the stroke records, the change situation mutation conditions of the target mediation index in different time spans in the mediation case database are evaluated, the characteristic mediation index is obtained, and the target time range of the characteristic mediation index for searching in the mediation case database is determined; And S4, analyzing the reference value generated by the adjustment record of the adjustment material of the historical adjustment cases in the adjustment case database according to the target time range, and assisting in the generation of the adjustment record of the adjustment material. Further, step S1 includes: step S11, acquiring a mediation material uploaded by a mediation person in a platform, and preprocessing the mediation material to obtain a mediation text in the mediation material; dividing the reconciliation text into sentences, extracting words in the sentences in the reconciliation text by using a natural language processing technology, and collecting the words to obtain a word set of the sentences; Step S12, acquiring a preset key index library, and acquiring a keyword set corresponding to each adjustment index from the key index library; acquiring word vectors of words in the word set of the sentence by using a word vector model preset by a platform, and calculat