CN117668221A - Vehicle fault handling mode pushing method, device and medium - Google Patents
Vehicle fault handling mode pushing method, device and medium Download PDFInfo
- Publication number
- CN117668221A CN117668221A CN202311362209.8A CN202311362209A CN117668221A CN 117668221 A CN117668221 A CN 117668221A CN 202311362209 A CN202311362209 A CN 202311362209A CN 117668221 A CN117668221 A CN 117668221A
- Authority
- CN
- China
- Prior art keywords
- vehicle fault
- text
- data
- repair data
- fault repair
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 61
- 230000008439 repair process Effects 0.000 claims abstract description 256
- 239000013598 vector Substances 0.000 claims abstract description 256
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 25
- 238000007621 cluster analysis Methods 0.000 claims abstract description 11
- 238000004590 computer program Methods 0.000 claims description 16
- 238000012545 processing Methods 0.000 claims description 13
- 235000002198 Annona diversifolia Nutrition 0.000 claims description 9
- 241000282842 Lama glama Species 0.000 claims description 9
- 238000009499 grossing Methods 0.000 claims description 9
- 238000007781 pre-processing Methods 0.000 claims description 9
- 230000011218 segmentation Effects 0.000 claims description 8
- 238000012935 Averaging Methods 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 abstract description 7
- 238000004891 communication Methods 0.000 description 8
- 230000006698 induction Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 238000012163 sequencing technique Methods 0.000 description 5
- 238000003672 processing method Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000013507 mapping Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 238000011524 similarity measure Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000009827 uniform distribution Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Abstract
The invention discloses a vehicle fault handling mode pushing method, equipment and medium. The method comprises the following steps: determining a second cosine similarity according to the first cosine similarity between any two text vectors corresponding to the historical vehicle fault repair data in a text vector set obtained by text description event conversion of the historical vehicle fault repair data; performing cluster analysis on the second cosine similarity of any two historical vehicle fault repair data based on a preset density clustering algorithm, and forming a text description event set of the historical vehicle fault repair data according to a clustering classification result of a preset large language model; and pushing solution data corresponding to the vehicle fault data to be treated based on matching of the historical vehicle fault report data text description event set and the vehicle fault data to be treated event. By the technical scheme, the problems of low efficiency and incomplete surface existing in practical application can be solved, and the query efficiency and accuracy of the new problem solution are improved.
Description
Technical Field
The invention relates to the technical field of vehicle fault handling, in particular to a vehicle fault handling mode pushing method, device and medium.
Background
With the continuous development of vehicle technology, vehicle faults are increased, and the vehicle fault disposal modes are increased, wherein the most important work for vehicle fault disposal is fault report repair and fault rush repair treatment work order circulation, so that a large number of event work orders are accumulated in daily maintenance work. There may be recurring or similar events in these large historical worksheets; in the prior art, a handling mode of a vehicle fault is usually implemented through a simple preprocessing operation, and then a keyword is directly searched between an newly added event to be processed and a stock event description text so as to push the handling mode of the vehicle fault, however, the mode does not carry out a certain early induction finishing operation on historical vehicle fault repair data, so that the method is difficult to accurately match with helpful past faults, and in the practical application, the problems that the efficiency of pushing the handling mode is extremely low and the pushed solution is not comprehensive and accurate exist.
Disclosure of Invention
In view of the above, the present invention provides a vehicle fault handling method, apparatus and medium for performing early induction and sorting on historical vehicle fault repair data, so as to accurately match a vehicle fault handling method corresponding to a historical fault, solve the problems of low efficiency and incomplete integrity in actual application, and improve the query efficiency and accuracy of a new problem solution.
According to an aspect of the present invention, an embodiment of the present invention provides a vehicle fault handling method, including:
acquiring a text description event of historical vehicle fault repair data, and converting the text description event of the historical vehicle fault repair data into a text vector form to obtain a corresponding text vector set;
for each text vector set, respectively determining first cosine similarity between corresponding message repair text vectors of any two historical vehicle fault repair data in each text vector set, and determining second cosine similarity of any two historical vehicle fault repair data according to each first cosine similarity;
performing cluster analysis on the second cosine similarity of the two random historical vehicle fault repair data based on a preset density clustering algorithm to obtain a clustering classification result corresponding to the text description event of the historical vehicle fault repair data, and forming the clustering classification result into a text description event set of the historical vehicle fault repair data according to a preset large language model;
acquiring a vehicle fault data event to be treated, and pushing solution data corresponding to the vehicle fault data problem to be treated based on matching of the historical vehicle fault report data text description event set and the vehicle fault data event to be treated.
According to another aspect of the present invention, an embodiment of the present invention further provides a vehicle fault handling manner pushing device, where the device includes:
the conversion module is used for acquiring a text description event of the historical vehicle fault repair data, converting the text description event of the historical vehicle fault repair data into a text vector form and obtaining a corresponding text vector set;
the similarity determining module is used for respectively determining first cosine similarity between the corresponding message repair text vectors of any two historical vehicle fault repair data in each text vector set according to each text vector set, and determining second cosine similarity of any two historical vehicle fault repair data according to each first cosine similarity;
the merging module is used for carrying out cluster analysis on the second cosine similarity of the fault repair data of any two historical vehicles based on a preset density clustering algorithm to obtain a cluster classification result corresponding to the text description event of the fault repair data of the historical vehicles, and forming the cluster classification result into a text description event set of the fault repair data of the historical vehicles according to a preset large language model;
and the pushing module is used for pushing the solution data corresponding to the vehicle fault data problem to be treated based on the matching of the historical vehicle fault report data text description event set and the vehicle fault data event to be treated.
According to another aspect of the present invention, an embodiment of the present invention further provides an electronic device, including:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,
the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the vehicle fault handling method according to any one of the embodiments of the present invention.
According to another aspect of the present invention, an embodiment of the present invention further provides a computer readable storage medium, where computer instructions are stored, where the computer instructions are configured to cause a processor to implement the vehicle fault handling method according to any one of the embodiments of the present invention when executed.
According to the technical scheme, text description events of the acquired historical vehicle fault repair data are converted into text vector forms, corresponding text vector sets are obtained, first cosine similarities between text vectors corresponding to any two historical vehicle fault repair data in each text vector set are respectively determined for each text vector set, second cosine similarities of any two historical vehicle fault repair data are determined according to the first cosine similarities, and robustness can be improved through the similarity of the fault text vectors; and carrying out cluster analysis on the second cosine similarity of any two historical vehicle fault repair data based on a preset density clustering algorithm to obtain a cluster classification result corresponding to the text description event of the historical vehicle fault repair data, forming a text description event set of the historical vehicle fault repair data according to a preset large language model, acquiring a vehicle fault data event to be treated, pushing solution data corresponding to the vehicle fault data problem to be treated based on matching of the text description event set of the historical vehicle fault repair data and the vehicle fault data event to be treated, and carrying out early induction finishing on the historical vehicle fault repair data, so that the vehicle fault treatment mode corresponding to the historical fault can be accurately matched, the problems of low efficiency and incomplete surface existing in practical application are solved, and the query efficiency and the accuracy of a new problem solution are improved.
It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the invention or to delineate the scope of the invention. Other features of the present invention will become apparent from the description that follows.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a vehicle fault handling pushing method according to an embodiment of the present invention;
FIG. 2 is a flowchart of another vehicle fault handling pushing method according to an embodiment of the present invention;
FIG. 3 is a block diagram illustrating a vehicle fault handling push device according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and the claims of the present invention and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the invention described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
In an embodiment, fig. 1 is a flowchart of a vehicle fault handling method according to an embodiment of the present invention, where the method may be performed by a vehicle fault handling method pushing device, and the vehicle fault handling method pushing device may be implemented in hardware and/or software, and the vehicle fault handling method pushing device may be configured in an electronic device.
As shown in fig. 1, the method for pushing the vehicle fault handling mode in this embodiment is applied to a database node, and the method specifically includes the following steps:
s110, acquiring a text description event of the historical vehicle fault repair data, and converting the text description event of the historical vehicle fault repair data into a text vector form to obtain a corresponding text vector set.
The historical vehicle fault repair data text description event can be understood as a processed historical vehicle fault repair data text description event. The text vector set refers to a text vector set obtained by text vectorizing a text description event of historical vehicle fault repair data, and the text vector set in the embodiment includes at least two text vectors.
In this embodiment, the text description event of the historical vehicle fault repair data at least includes: historical vehicle fault repair data, solution data corresponding to the historical vehicle fault repair data, vehicle repair type, and vehicle number. In this embodiment, the fault repair data corresponds to the solution data one by one, and if n pieces of fault repair data exist, n pieces of solution data exist.
In this embodiment, a historical vehicle fault repair data text description event may be obtained from a service system, data preprocessing is performed on the historical vehicle fault repair data text description event to obtain a preprocessed target vehicle fault repair data text description event, a plurality of data preprocessing modes may include, for the target vehicle fault repair data text description event, a plurality of preset text vector methods may be adopted to convert the target vehicle fault repair data text description event into a text vector form, and a corresponding text vector set is obtained, in this embodiment, each preset text vector method corresponds to one text vector set, and the expression forms of text vectors in each text vector set are different; each text vector represents a text descriptive event for historical vehicle fault repair data.
S120, for each text vector set, determining first cosine similarity between corresponding message repair text vectors of any two historical vehicle fault repair data in each text vector set respectively, and determining second cosine similarity of any two historical vehicle fault repair data according to each first cosine similarity.
The first cosine similarity refers to cosine similarity between message text vectors corresponding to any two historical vehicle fault message data in the text vector set. The second cosine similarity refers to the cosine similarity obtained by performing mean value operation on each first cosine similarity.
In this embodiment, vector smoothing may be performed on the repair text vectors corresponding to any two historical vehicle fault repair data in each text vector set to obtain a smoothed text vector set, so as to reduce vector sparsity through text vector smoothing, determine that any two historical vehicle fault repair data in the smoothed text vector set correspond to the repair text vectors respectively, determine first cosine similarity between the repair text vectors according to a preset first cosine similarity formula, and use an average value of the first cosine similarity as second cosine similarity of any two historical vehicle fault repair data; in some embodiments, the cosine similarity between the text vectors corresponding to any two historical vehicle fault repair data in each text vector set may also be determined by a multidimensional cosine similarity formula, which is not limited herein.
S130, carrying out cluster analysis on the second cosine similarity of any two historical vehicle fault repair data based on a preset density clustering algorithm to obtain a cluster classification result corresponding to the text description event of the historical vehicle fault repair data, and forming the cluster classification result into a text description event set of the historical vehicle fault repair data according to a preset large language model.
The historical vehicle fault repair data text description event set comprises a plurality of historical vehicle fault repair data text description event subsets, wherein each subset corresponds to the same type of fault repair data, namely, each subset corresponds to one type.
In this embodiment, the preset density clustering algorithm may include, but is not limited to, a DBSCAN classification and an OPTICS clustering algorithm, and may perform induction sorting on fault repair data and solution data according to a result of similarity calculation, that is, text with high similarity is classified into one type, in this embodiment, cluster analysis is performed on second cosine similarity of any two historical vehicle fault repair data to obtain a cluster classification result corresponding to a text description event of the historical vehicle fault repair data, and specifically, two parameters of the DBSCAN algorithm are set, namely, a neighborhood radius Eps and a minimum point number MinPts of a core object; from any text vector as a starting point, the number of neighborhood texts in the Eps range is found. If the number is greater than or equal to MinPts, forming a cluster; if less, the mark is noise; repeating the steps for the remaining unvisited text vectors; until all text vectors are accessed; finally, all the text vectors connected in density are divided into the same cluster, and points which cannot be connected to a high-density area are marked as noise; and obtaining clustering division of the text set according to the density clustering result of the text vector, and obtaining the report repair data classification.
In an embodiment, the preset large language model may be one of ChatGLM, baichuan, llama, and the understanding and generating capabilities of the large language model (one of ChatGLM, baichuan, llama) may be utilized to sort the same type of repair data and solutions, where in this embodiment, the text description events of historical vehicle fault repair data with similar or same type in the clustered classification result may be combined to obtain a first text description event subset of vehicle fault repair data, and other non-similar text description events of vehicle fault repair data may be listed in a striped manner to obtain multiple text description event subsets of vehicle fault repair data; in some embodiments, the plurality of vehicle fault repair data text description event subsets may also be obtained by combining the historical vehicle fault repair data text description events having similar or same type in the cluster classification result in other manners, which is not limited herein.
S140, acquiring a vehicle fault data event to be treated, and pushing solution data corresponding to the vehicle fault data problem to be treated based on matching of the historical vehicle fault report data text description event set and the vehicle fault data event to be treated.
The preset TextRank algorithm refers to a keyword extraction and summary algorithm. The first text vector may be understood as a text vector corresponding to a keyword for which the vehicle fault data event is to be handled. The second text vector refers to the text vector corresponding to each historical vehicle fault repair data text description event in the historical vehicle fault repair data text description event subset matched with the keyword. The third cosine similarity refers to cosine similarity between the first text vector and the second text vector.
In this embodiment, according to a new fault problem input by a user, a TextRank algorithm is used to obtain a fault problem keyword, all fault data sets including the keyword are obtained from fault repair data which are already summarized, a first text vector corresponding to the keyword is obtained through any one of Sentence Transformer, roBERTa-WWM and Doc2vec, a second text vector corresponding to a text description event subset of historical vehicle fault repair data matched with the keyword is obtained, vectorized cosine similarity is calculated, similarity matching is performed, and when a threshold is met, a solution corresponding to the most similar fault data is returned as a reference answer. Specifically, word segmentation is carried out on text data in a vehicle fault data event to be treated to obtain candidate keywords, an undirected weighted graph among each word in the candidate keywords is constructed, textnodenrank of each word is obtained according to the undirected weighted graph to obtain keywords, a text description event subset of target vehicle fault repair data of keywords contained in a historical vehicle fault repair data set is matched according to the keywords, vectorized cosine similarity is calculated, similarity matching is carried out, and when a preset cosine similarity threshold is met, a solution corresponding to the most similar fault data is returned as a reference answer.
According to the technical scheme, text description events of the acquired historical vehicle fault repair data are converted into text vector forms, corresponding text vector sets are obtained, first cosine similarities between text vectors corresponding to any two historical vehicle fault repair data in each text vector set are respectively determined for each text vector set, second cosine similarities of any two historical vehicle fault repair data are determined according to the first cosine similarities, and robustness can be improved through the similarity of the fault text vectors; and carrying out cluster analysis on the second cosine similarity of any two historical vehicle fault repair data based on a preset density clustering algorithm to obtain a cluster classification result corresponding to the text description event of the historical vehicle fault repair data, forming a text description event set of the historical vehicle fault repair data according to a preset large language model, acquiring a vehicle fault data event to be treated, pushing solution data corresponding to the vehicle fault data problem to be treated based on matching of the text description event set of the historical vehicle fault repair data and the vehicle fault data event to be treated, and carrying out early induction finishing on the historical vehicle fault repair data, so that the vehicle fault treatment mode corresponding to the historical fault can be accurately matched, the problems of low efficiency and incomplete surface existing in practical application are solved, and the query efficiency and the accuracy of a new problem solution are improved.
In an embodiment, fig. 2 is a flowchart of another vehicle fault handling method according to an embodiment of the present invention, where on the basis of the foregoing embodiments, a text description event of historical vehicle fault repair data is converted into a text vector form, a corresponding text vector set is obtained, first cosine similarities between text vectors corresponding to any two historical vehicle fault repair data in each text vector set are respectively determined, second cosine similarities of the any two historical vehicle fault repair data are determined according to each first cosine similarity, a clustering classification result corresponding to the text description event of the historical vehicle fault repair data is obtained by performing a clustering analysis on the second cosine similarities of the any two historical vehicle fault repair data based on a preset density clustering algorithm, and the clustering classification result is formed into a text description event set of the historical vehicle fault repair data according to a preset large language model, and solution data corresponding to the vehicle fault data to be handled is further refined based on matching of the text description event set of the historical vehicle fault repair data and the vehicle fault data to be handled.
As shown in fig. 2, the vehicle fault handling method in this embodiment may specifically include the following steps:
s210, acquiring a historical vehicle fault repair data text description event, and performing data preprocessing on the historical vehicle fault repair data text description event to obtain a preprocessed target vehicle fault repair data text description event.
In this embodiment, the data preprocessing is performed on the text description event of the historical vehicle fault repair data to obtain the preprocessed text description event of the target vehicle fault repair data, which may be understood as performing data cleaning on the fault repair data and the solution data, where the data preprocessing mode includes at least one of the following: filtering text description events of historical vehicle fault repair data; removing English and digital contents in text description events of historical vehicle fault report data; and mapping the field names of the vehicles in the text description event of the historical vehicle fault report and repair data, and uniformly mapping the vehicles of the same type. Exemplary, filtering meaningless message text descriptions such as o's, emoticons; removing irrelevant contents such as English, numbers and the like in the text; correctly mapping the product name field; (e.g., the correct product name is T, which is referred to in the marketplace as T1, T2, T3, then uniformly mapped to T).
S220, converting the text description event of the fault report data of the target vehicle into a text vector form by adopting at least three preset text vector methods to obtain at least three corresponding text vector sets.
The method for presetting the text vector at least comprises the following steps: sentence Transformer, roberta-WWM and Doc2vec.
In this embodiment, at least three preset text vector methods may be adopted to convert the text description event of the fault report data of the target vehicle into a text vector form, and convert the text feature into a numerical vector, so as to obtain at least three corresponding text vector sets, where at least three text vector sets respectively include at least two text vectors; the expression forms of the text vectors in the text vector sets are different; each text vector represents a text descriptive event for historical vehicle fault repair data. For example, three modes of Sentence Transformer, roberta-WWM and Doc2vec are used for text vectorization, so that 3 text vector sets A0, B0 and C0 are respectively obtained, each set contains n sentence vectors, and each text vector represents one piece of fault repair data.
S230, carrying out vector smoothing processing on the message repair text vectors corresponding to any two historical vehicle fault repair data in each text vector set aiming at each text vector set to obtain a smoothed text vector set.
In this embodiment, for each text vector set, vector smoothing is performed on the repair text vectors corresponding to any two historical vehicle fault repair data in each text vector set to obtain a smoothed text vector set, which can be understood that after outputting the text vector, a tiny random value or noise (randomly sampled from a uniform distribution [ -1e-9,1 e-9) is added to the text vector to implement vector smoothing, and smoothed text vector sets a, B, and C are obtained.
S240, determining first cosine similarity between any two historical vehicle fault repair data in the smoothed text vector set according to a preset first cosine similarity formula, wherein the first cosine similarity corresponds to the repair text vectors.
In this embodiment, the first cosine similarity between any two historical vehicle fault repair data in the smoothed text vector set is determined according to a preset first cosine similarity formula, where the two historical vehicle fault repair data correspond to the repair text vectors respectively. The method comprises the steps of presetting a first cosine similarity formula, wherein the formula is expressed as: fs= (fa1·fa2)/(|fa1|fa2|); wherein, FA1 and FA2 are respectively expressed as any two historical vehicle fault report data which respectively correspond to report text vectors; the I FA1 and the I FA2 are respectively expressed as modes of any two historical vehicle fault repair data corresponding to the repair text vectors respectively.
For example, for any two fault repair data F1, F2, after the fault repair data are respectively calculated and converted into text vectors, cosine similarity of the text vectors corresponding to F1, F2 in the set of a, B, C is calculated. For example, in the set a, F1 corresponds to the text vector FA1, F2 corresponds to the text vector FA2, cosine similarity is calculated, and the dot product is divided by the product of two vector modes, as follows: fas= (fa1·fa2)/(|fa1|fa2|); the cosine similarity FBS of F1, F2 in the B set and the cosine similarity FCS of F1, F2 in the C set are obtained by the same method.
S250, averaging the first cosine similarities to obtain second cosine similarities of any two historical vehicle fault repair data.
Wherein the second cosine similarity has a value range of [ -1,1].
In this embodiment, the second cosine similarity of any two historical vehicle fault repair data is obtained by taking the average value of the first cosine similarities. The mean value of three cosine similarities is taken as the similarity measure of F1 and F2: similarity (F1, F2) =similarity (F2, F1) = (fas+fbs+fcs)/3. The similarity value range is [ -1,1].
S260, randomly selecting a text vector corresponding to the second cosine similarity from the second cosine similarities as a current text vector, judging whether the current text vector has a preset number of neighborhood text vectors within a preset distance threshold, if so, executing S270, and if not, executing S280.
Wherein, the preset distance threshold value is expressed as the following formula: 1-Similarity, wherein the Similarity is the second cosine Similarity of any two vehicle fault repair data; the value range of the preset distance threshold is expressed as follows: the smaller the value is, the higher the second cosine similarity of any two historical vehicle fault report repair data is represented; and the clustering classification result corresponds to the corresponding classification label.
In this embodiment, a text vector corresponding to a second cosine similarity is randomly selected from the second cosine similarities as a current text vector, whether a preset number of neighborhood text vectors exist in the current text vector within a preset distance threshold is judged, if yes, the current text vector is used as a core point, other text vectors of the core point within the preset distance threshold are uniformly set as the core point, a core point set is formed as a cluster, a next text vector is used as the current text vector, and the step of judging whether the preset number of neighborhood text vectors exist in the current text vector within the preset distance threshold is returned until all the text vectors corresponding to the second cosine similarities are completely divided, so that a clustering classification result is obtained; if not, marking the current text vector as noise, wherein the noise is used as one of the clustering division results.
S270, taking the current text vector as a core point, uniformly setting other text vectors of the core point within a preset distance threshold as core points, forming a core point set as a cluster, taking the next text vector as the current text vector, and returning to the step of judging whether a preset number of neighborhood text vectors exist in the current text vector within the preset distance threshold until all text vectors corresponding to the second cosine similarity are completely divided, thereby obtaining a clustering classification result.
In this embodiment, when a preset number of neighbor text vectors exist in the current text vector within the preset distance threshold, the current text vector is used as a core point, other text vectors of the core point within the preset distance threshold are uniformly set as the core point, a core point set is formed as a cluster, the next text vector is used as the current text vector, and the step of judging whether the current text vector has the preset number of neighbor text vectors within the preset distance threshold is returned until all text vectors corresponding to the second cosine similarity are divided, so as to obtain a clustering classification result.
S280, marking the current text vector as noise, wherein the noise is used as one of the clustering division results.
In this embodiment, if the current text vector does not have a preset number of neighboring text vectors within a preset distance threshold, the current text vector is marked as noise, where the noise is used as one of the cluster division results.
And S290, merging the text description events with similar or same type of historical vehicle fault repair data in the clustering classification result according to any one of ChatGLM, baichuan or Llama to obtain a first vehicle fault repair data text description event subset, and carrying out strip listing on other non-similar vehicle fault repair data text description events to serve as a second vehicle fault repair data text description event subset.
Wherein each second vehicle fault repair data text description event subset corresponds to a vehicle fault repair type.
In this embodiment, the same type of repair data and solutions are consolidated using the understanding and generating capabilities of the large language model (one of ChatGLM, baichuan, llama). Specifically, according to any one of ChatGLM, baichuan or Llama, the text description events with similar or same type of historical vehicle fault repair data in the clustering classification result are combined to obtain a first vehicle fault repair data text description event subset, and other non-similar vehicle fault repair data text description events are listed in a split manner to be used as a second vehicle fault repair data text description event subset.
And S2100, forming the first vehicle failure repair data text description event subset and the second vehicle failure repair data text description event subset into the historical vehicle failure repair data text description event set.
In this embodiment, the first vehicle failure repair data text description event subset and the second vehicle failure repair data text description event subset are formed into the historical vehicle failure repair data text description event set.
S2110, acquiring a vehicle fault data event to be treated, and determining keywords of the vehicle fault data event to be treated according to a preset TextRank algorithm.
The candidate keywords comprise all words in the text data after word segmentation.
In this embodiment, a vehicle fault data event to be handled is acquired, and text data in the vehicle fault data event to be handled is subjected to word segmentation processing to obtain candidate keywords. Specifically, text data in a fault data event of a vehicle to be treated can be subjected to word segmentation to obtain candidate keywords, and the words in the range of a preset ranking threshold value are selected to be ranked to reach the set threshold value or directly taken as the keywords according to the candidate keywords and the preset ranking.
In an embodiment, determining keywords of a vehicle fault data event to be handled according to a preset TextRank algorithm includes: word segmentation processing is carried out on text data in a vehicle fault data event to be treated to obtain candidate keywords;
constructing an undirected weighted graph among each word in the candidate keywords, and obtaining textnodenRank of each word according to the undirected weighted graph;
and sequencing the words according to the textNodeRank, and selecting words with the ranking reaching a set threshold value in the sequencing or directly taking the words within the range of a preset ranking threshold value as key words.
The edge weight among the words can be obtained by using the co-occurrence frequency of the words or the relativity of the words. The preset ranking threshold range may be set by a person by himself, generally taking the top 3, which is not limited in this embodiment.
In this embodiment, an undirected weighted graph between words is constructed, and iterative computation is performed on the graph to obtain an importance score (textnodank) of each word, where the edge weights between the words may use co-occurrence frequency of the words or word relativity. The words are ordered according to the textnodenrank, the words with the top ranking are selected as keywords, or a threshold value can be set, and the words with the textnodenrank larger than the threshold value are selected as keywords. The word of the top 3 rank can also be directly taken as the keyword, and the extracted keyword is returned. And matching the keywords with a target vehicle fault repair data text description event subset of keywords contained in the historical vehicle fault repair data set, and obtaining all fault data sets containing the keywords.
S2120, pushing solution data corresponding to the to-be-treated vehicle fault data problem according to the first text vector corresponding to the keyword and third cosine similarity between the text description event subset of the historical vehicle fault repair data matched with the keyword in the text description event set of the historical vehicle fault repair data.
In this embodiment, solution data corresponding to the to-be-handled vehicle fault data problem may be pushed according to the first text vector corresponding to the keyword and the third cosine similarity between the second text vector corresponding to the subset of the historical vehicle fault repair data text description event set matched with the keyword in the historical vehicle fault repair data text description event set.
In an embodiment, pushing solution data corresponding to the to-be-treated vehicle fault data problem according to a first text vector corresponding to the keyword and a third cosine similarity between a text description event subset of historical vehicle fault repair data matched with the keyword in the text description event set of the historical vehicle fault repair data, where the third cosine similarity is between the text description event subset of the historical vehicle fault repair data and the second text vector, includes:
describing an event subset according to a target vehicle fault repair data text of which the keywords are matched with keywords contained in the historical vehicle fault repair data set;
Determining third cosine similarity between a second text vector corresponding to each historical vehicle fault repair data text description event in the target vehicle fault repair data text description event subset and a first text vector corresponding to the keyword according to a preset second cosine similarity formula;
and pushing solution data corresponding to the to-be-treated vehicle fault data problem according to the third cosine similarity and the preset cosine similarity threshold.
The preset cosine similarity threshold may be set by experience and manually, which is not limited in this embodiment.
In this embodiment, according to the target vehicle fault repair data text description event subset of the keywords included in the keyword matching historical vehicle fault repair data set, determining third cosine similarity between the second text vector corresponding to each historical vehicle fault repair data text description event in the target vehicle fault repair data text description event subset and the first text vector corresponding to the keyword according to a preset second cosine similarity formula, and pushing solution data corresponding to the to-be-treated vehicle fault data problem according to the third cosine similarity and a preset cosine similarity threshold. Specifically, under the condition that the third cosine similarity meets the preset cosine similarity threshold, returning a solution corresponding to the most similar fault data as a reference answer. It should be noted that, the preset second cosine similarity formula is consistent with the preset first cosine similarity formula calculation method, and this embodiment is not separately described herein.
According to the technical scheme, the data is preprocessed to obtain the preprocessed target vehicle fault repair data text description event, at least three preset text vector methods are adopted to convert the target vehicle fault repair data text description event into a text vector form, at least three corresponding text vector sets are obtained, the similarity of fault text vectors obtained by using an integrated learning method is achieved, and the robustness is improved; carrying out vector smoothing on message repair text vectors corresponding to any two historical vehicle fault repair data in each text vector set to obtain a smoothed text vector set, so as to realize text vector smoothing and reduce vector sparsity; according to a preset density clustering algorithm, clustering analysis is carried out on the second cosine similarity of the random two historical vehicle fault repair data to obtain a clustering classification result corresponding to the historical vehicle fault repair data text description event, merging historical vehicle fault repair data text description events with similar or same type in the clustering classification result according to any one of ChatGLM, baichuan or Llama to obtain a first vehicle fault repair data text description event subset, carrying out stripe listing on other dissimilar vehicle fault repair data text description events to obtain a vehicle fault data event to be treated, determining a keyword of the vehicle fault data event to be treated according to a preset TextRank algorithm, carrying out clustering analysis on the first text vector corresponding to the first historical vehicle fault repair data text description event to obtain a clustering classification result corresponding to the historical vehicle fault repair data text description event, carrying out merging on the historical vehicle fault repair data text description event with similar or same type in the clustering classification result according to any one of ChatGLM, baichuan or Llama mode to obtain a first vehicle fault repair data text description event subset, carrying out stripe listing on the other dissimilar vehicle fault repair data text description events, and carrying out overall solving the problem solving when the problem of the corresponding to the first vehicle fault repair data text description event is not matched with the corresponding to the historical text description event text description word of the actual text description word of the historical text, and the actual fault repair data is completely solved.
In an embodiment, fig. 3 is a block diagram of a vehicle fault handling pushing device according to an embodiment of the present invention, where the device is applicable to a situation when performing fault handling pushing on a vehicle fault, and the device may be implemented by hardware/software. The vehicle fault handling push processing method can be configured in the electronic equipment to realize the vehicle fault handling push processing method in the embodiment of the invention.
As shown in fig. 3, the apparatus includes: the device comprises a conversion module 310, a similarity determination module 320, a combination module 330 and a pushing module 340.
The conversion module 310 is configured to obtain a text description event of the historical vehicle fault repair data, and convert the text description event of the historical vehicle fault repair data into a text vector form, so as to obtain a corresponding text vector set;
the similarity determining module 320 is configured to determine, for each text vector set, a first cosine similarity between text vectors corresponding to any two pieces of historical vehicle fault repair data in each text vector set, and determine a second cosine similarity of any two pieces of historical vehicle fault repair data according to each first cosine similarity;
the merging module 330 is configured to perform cluster analysis on the second cosine similarity of the two pieces of historical vehicle fault repair data based on a preset density clustering algorithm to obtain a cluster classification result corresponding to the text description event of the historical vehicle fault repair data, and form the cluster classification result into a text description event set of the historical vehicle fault repair data according to a preset large language model;
The pushing module 340 is configured to obtain a vehicle fault data event to be handled, and push solution data corresponding to the vehicle fault data problem to be handled based on matching of the historical vehicle fault repair data text description event set and the vehicle fault data event to be handled.
According to the embodiment of the invention, the text description event of the acquired historical vehicle fault repair data is converted into a text vector form through the conversion module, a corresponding text vector set is obtained, and the similarity determination module is used for respectively determining the first cosine similarity between the text vectors of the corresponding message of any two historical vehicle fault repair data in each text vector set according to each text vector set, and determining the second cosine similarity of any two historical vehicle fault repair data according to each first cosine similarity, so that the robustness can be improved through the similarity of the fault text vectors; the merging module performs cluster analysis on the second cosine similarity of any two historical vehicle fault repair data based on a preset density clustering algorithm to obtain a cluster classification result corresponding to a text description event of the historical vehicle fault repair data, composes the cluster classification result into a text description event set of the historical vehicle fault repair data according to a preset large language model, acquires the vehicle fault data event to be treated, pushes solution data corresponding to the vehicle fault data problem to be treated based on matching of the text description event set of the historical vehicle fault repair data and the vehicle fault data event to be treated, and performs early induction and finishing on the historical vehicle fault repair data, so that the vehicle fault treatment mode corresponding to the historical fault can be accurately matched, the problems of low efficiency and incomplete surface existing in practical application are solved, and the query efficiency and the accuracy of a new problem solution are improved.
In one embodiment, the conversion module 310 includes:
the preprocessing unit is used for preprocessing the data of the historical vehicle fault report data text description event to obtain a preprocessed target vehicle fault report data text description event;
the conversion unit is used for converting the text description event of the fault report data of the target vehicle into a text vector form by adopting at least three preset text vector methods to obtain at least three corresponding text vector sets; the preset text vector method at least comprises the following steps: sentence Transformer, roBERTa-WWM and Doc2vec;
wherein the at least three text vector sets respectively comprise at least two text vectors; the expression forms of the text vectors in the text vector sets are different; each text vector represents a text descriptive event for historical vehicle fault repair data.
In one embodiment, the similarity determination module 320 includes:
the processing unit is used for carrying out vector smoothing processing on the message text vectors corresponding to any two pieces of historical vehicle fault report data in each text vector set to obtain a smoothed text vector set;
the first determining unit is used for determining first cosine similarity between any two historical vehicle fault repair data in the smoothed text vector set respectively corresponding to the repair text vectors according to a preset first cosine similarity formula; the first cosine similarity formula is preset, and the formula is expressed as: fs= (fa1·fa2)/(|fa1|fa2|); the FA1 and the FA2 are respectively expressed as the repair text vectors corresponding to the random two historical vehicle fault repair data; the |FA1| and the |FA2| are respectively expressed as models of the repair text vectors corresponding to the fault repair data of any two historical vehicles respectively;
And the second determining unit is used for averaging the first cosine similarities to obtain second cosine similarities of the any two historical vehicle fault repair data.
In one embodiment, the merging module 330 includes:
the judging unit is used for randomly selecting a text vector corresponding to the second cosine similarity from the second cosine similarity as a current text vector and judging whether a preset number of neighborhood text vectors exist in a preset distance threshold value of the current text vector;
the first clustering unit is used for taking the current text vector as a core point, uniformly setting other text vectors of the core point as the core point within a preset distance threshold value to form a core point set as a cluster, taking the next text vector as the current text vector, and returning to the step of judging whether a preset number of neighborhood text vectors exist in the current text vector within the preset distance threshold value until all text vectors corresponding to the second cosine similarity are completely divided, so as to obtain a clustering classification result;
and the first clustering unit is used for marking the current text vector as noise if the current text vector does not exist, wherein the noise is used as one of the clustering division results.
In one embodiment, the merging module 330 further includes:
the merging unit is used for merging the text description events with similar or same type of historical vehicle fault repair data in the clustering classification result according to any one of ChatGLM, baichuan or Llama to obtain a first vehicle fault repair data text description event subset, and carrying out strip listing on other non-similar vehicle fault repair data text description events to serve as a second vehicle fault repair data text description event subset; wherein each second vehicle fault repair data text description event subset corresponds to a vehicle fault repair type;
the composition unit is used for composing the first vehicle fault repair data text description event subset and the second vehicle fault repair data text description event subset into the historical vehicle fault repair data text description event set.
In one embodiment, the pushing module 340 includes:
the determining unit is used for determining keywords of the vehicle fault data event to be treated according to a preset TextRank algorithm;
and the pushing unit is used for pushing the solution data corresponding to the to-be-treated vehicle fault data problem according to the first text vector corresponding to the keyword and the third cosine similarity between the second text vectors corresponding to the text description event subset of the historical vehicle fault repair data matched with the keyword in the text description event set of the historical vehicle fault repair data.
In an embodiment, the determining unit comprises:
the processing subunit is used for word segmentation processing of the text data in the vehicle fault data event to be treated to obtain candidate keywords; the candidate keywords comprise all words in text data after word segmentation;
the construction subunit is used for constructing an undirected weighted graph among each word in the candidate keywords and obtaining the textnodenRank of each word according to the undirected weighted graph;
and the sequencing subunit is used for sequencing the words according to the textnodenrank, and selecting words in the sequencing, the ranking of which reaches a set threshold value or directly taking words in a preset ranking threshold value range as the keywords.
In an embodiment, the pushing unit comprises:
the matching subunit is used for matching the target vehicle fault repair data text description event subset of the keywords contained in the historical vehicle fault repair data set according to the keywords;
the determining subunit is configured to determine a third cosine similarity between a second text vector corresponding to each historical vehicle fault repair data text description event in the target vehicle fault repair data text description event subset and a first text vector corresponding to the keyword according to a preset second cosine similarity formula;
And the pushing subunit is used for pushing the solution data corresponding to the to-be-treated vehicle fault data problem according to the third cosine similarity and a preset cosine similarity threshold.
The vehicle fault handling mode pushing device provided by the embodiment of the invention can execute the vehicle fault handling mode pushing processing method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the executing method.
In an embodiment, fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present invention. The electronic device 10 is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. Electronic equipment may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the inventions described and/or claimed herein.
As shown in fig. 4, the electronic device 10 includes at least one processor 11, and a memory, such as a Read Only Memory (ROM) 12, a Random Access Memory (RAM) 13, etc., communicatively connected to the at least one processor 11, in which the memory stores a computer program executable by the at least one processor, and the processor 11 may perform various appropriate actions and processes according to the computer program stored in the Read Only Memory (ROM) 12 or the computer program loaded from the storage unit 18 into the Random Access Memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 may also be stored. The processor 11, the ROM 12 and the RAM 13 are connected to each other via a bus 14. An input/output (I/O) interface 15 is also connected to bus 14.
Various components in the electronic device 10 are connected to the I/O interface 15, including: an input unit 16 such as a keyboard, a mouse, etc.; an output unit 17 such as various types of displays, speakers, and the like; a storage unit 18 such as a magnetic disk, an optical disk, or the like; and a communication unit 19 such as a network card, modem, wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunication networks.
The processor 11 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various processors running machine learning model algorithms, digital Signal Processors (DSPs), and any suitable processor, controller, microcontroller, etc. The processor 11 performs the various methods and processes described above, such as the vehicle fault handling push method.
In some embodiments, the vehicle fault handling push processing method may be implemented as a computer program tangibly embodied on a computer-readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program may be loaded and/or installed onto the electronic device 10 via the ROM 12 and/or the communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the vehicle fault handling push method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the vehicle fault handling mode push method by any other suitable means (e.g., by means of firmware).
Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.
A computer program for carrying out methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general purpose computer, special purpose computer, or other programmable vehicle fault handling means, such that the computer programs, when executed by the processor, cause the functions/operations specified in the flowchart and/or block diagram block or blocks to be implemented. The computer program may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.
In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, the computer readable storage medium may be a machine readable signal medium. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) through which a user can provide input to the electronic device. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), blockchain networks, and the internet.
The computing system may include clients and servers. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical hosts and VPS service are overcome.
It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps described in the present invention may be performed in parallel, sequentially, or in a different order, so long as the desired results of the technical solution of the present invention are achieved, and the present invention is not limited herein.
The above embodiments do not limit the scope of the present invention. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of the present invention.
Claims (10)
1. A vehicle fault handling method push method, comprising:
acquiring a text description event of historical vehicle fault repair data, and converting the text description event of the historical vehicle fault repair data into a text vector form to obtain a corresponding text vector set;
for each text vector set, respectively determining first cosine similarity between corresponding message repair text vectors of any two historical vehicle fault repair data in each text vector set, and determining second cosine similarity of any two historical vehicle fault repair data according to each first cosine similarity;
Performing cluster analysis on the second cosine similarity of the two random historical vehicle fault repair data based on a preset density clustering algorithm to obtain a clustering classification result corresponding to the text description event of the historical vehicle fault repair data, and forming the clustering classification result into a text description event set of the historical vehicle fault repair data according to a preset large language model;
acquiring a vehicle fault data event to be treated, and pushing solution data corresponding to the vehicle fault data problem to be treated based on matching of the historical vehicle fault report data text description event set and the vehicle fault data event to be treated.
2. The method of claim 1, wherein converting the historical vehicle repair data text description event into text vector form, resulting in a corresponding set of text vectors, comprises:
performing data preprocessing on the historical vehicle fault report repair data text description event to obtain a preprocessed target vehicle fault report repair data text description event;
converting the text description event of the fault report data of the target vehicle into a text vector form by adopting at least three preset text vector methods to obtain at least three corresponding text vector sets; the preset text vector method at least comprises the following steps: sentence Transformer, roBERTa-WWM and Doc2vec;
Wherein the at least three text vector sets respectively comprise at least two text vectors; the expression forms of the text vectors in the text vector sets are different; each text vector represents a text descriptive event for historical vehicle fault repair data.
3. The method of claim 1, wherein the determining the first cosine similarity between any two of the text vector sets and the corresponding repair text vectors of the historical vehicle fault repair data, and determining the second cosine similarity of any two of the historical vehicle fault repair data according to the first cosine similarity, respectively, comprises:
vector smoothing is carried out on the message repair text vectors corresponding to any two historical vehicle fault repair data in each text vector set to obtain a smoothed text vector set;
determining first cosine similarity between any two historical vehicle fault repair data in the smoothed text vector set according to a preset first cosine similarity formula; the first cosine similarity formula is preset, and the formula is expressed as: fs= (fa1·fa2)/(|fa1|fa2|); the FA1 and the FA2 are respectively expressed as the repair text vectors corresponding to the random two historical vehicle fault repair data; the |FA1| and the |FA2| are respectively expressed as models of the repair text vectors corresponding to the fault repair data of any two historical vehicles respectively;
And averaging the first cosine similarities to obtain second cosine similarities of the fault repairing data of any two historical vehicles.
4. The method of claim 1, wherein the performing cluster analysis on the second cosine similarity of the any two historical vehicle fault repair data based on the preset density clustering algorithm to obtain a cluster classification result corresponding to the text description event of the historical vehicle fault repair data includes:
randomly selecting a text vector corresponding to the second cosine similarity from the second cosine similarity as a current text vector, and judging whether a preset number of neighborhood text vectors exist in a preset distance threshold of the current text vector;
if so, uniformly setting the current text vector as a core point, uniformly setting other text vectors of the core point within a preset distance threshold as the core point to form a core point set as a cluster, taking the next text vector as the current text vector, and returning to the step of judging whether a preset number of neighborhood text vectors exist in the current text vector within the preset distance threshold until all text vectors corresponding to the second cosine similarity are completely divided, thereby obtaining a clustering classification result;
And if the current text vector does not exist, marking the current text vector as noise, wherein the noise is used as one of the clustering division results.
5. The method of claim 1, wherein the grouping the cluster classification results into a set of historical vehicle repair failure data text description events according to a pre-set large language model comprises:
combining the text description events with similar or same type of historical vehicle fault repair data in the clustering classification result according to any one mode of ChatGLM, baichuan or Llama to obtain a first vehicle fault repair data text description event subset, and carrying out strip listing on other non-similar vehicle fault repair data text description events to serve as a second vehicle fault repair data text description event subset; wherein each second vehicle fault repair data text description event subset corresponds to a vehicle fault repair type;
and forming the first vehicle failure repair data text description event subset and the second vehicle failure repair data text description event subset into the historical vehicle failure repair data text description event set.
6. The method of claim 1, wherein the pushing solution data corresponding to the vehicle fault data problem to be treated based on the matching of the set of historical vehicle fault repair data text description events to the vehicle fault data event to be treated comprises:
Determining keywords of the vehicle fault data event to be treated according to a preset TextRank algorithm;
and pushing the solution data corresponding to the to-be-treated vehicle fault data problem according to the first text vector corresponding to the keyword and the third cosine similarity between the text description event subset of the historical vehicle fault repair data matched with the keyword in the text description event set of the historical vehicle fault repair data.
7. The method of claim 6, wherein the determining keywords of the vehicle fault data event to be handled according to a preset TextRank algorithm comprises:
word segmentation processing is carried out on text data in the vehicle fault data event to be treated to obtain candidate keywords;
constructing an undirected weighted graph among each word in the candidate keywords, and obtaining the textnodenRank of each word according to the undirected weighted graph;
and sorting the words according to the textNodeRank, and selecting words in the sorting, the ranking of which reaches a set threshold value, or directly taking the words in the range of a preset ranking threshold value as the keywords.
8. The method of claim 6, wherein pushing the solution data corresponding to the vehicle fault data to be treated according to the first text vector corresponding to the keyword and a third cosine similarity between a second text vector corresponding to a subset of historical vehicle fault repair data text description events in the set of historical vehicle fault repair data text description events that match the keyword, comprises:
According to the target vehicle fault repair data text description event subset of which the keywords are matched with the keywords in the historical vehicle fault repair data set;
determining third cosine similarity between a second text vector corresponding to each historical vehicle fault repair data text description event in the target vehicle fault repair data text description event subset and a first text vector corresponding to the keyword according to a preset second cosine similarity formula;
and pushing solution data corresponding to the to-be-treated vehicle fault data problem according to the third cosine similarity and a preset cosine similarity threshold.
9. An electronic device, the electronic device comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,
the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the vehicle fault handling method of any one of claims 1-8.
10. A computer readable storage medium storing computer instructions for causing a processor to implement the vehicle fault handling mode pushing method of any one of claims 1-8 when executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311362209.8A CN117668221A (en) | 2023-10-19 | 2023-10-19 | Vehicle fault handling mode pushing method, device and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311362209.8A CN117668221A (en) | 2023-10-19 | 2023-10-19 | Vehicle fault handling mode pushing method, device and medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117668221A true CN117668221A (en) | 2024-03-08 |
Family
ID=90081473
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311362209.8A Pending CN117668221A (en) | 2023-10-19 | 2023-10-19 | Vehicle fault handling mode pushing method, device and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117668221A (en) |
-
2023
- 2023-10-19 CN CN202311362209.8A patent/CN117668221A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108804641B (en) | Text similarity calculation method, device, equipment and storage medium | |
| CN111797210A (en) | Information recommendation method, device and equipment based on user portrait and storage medium | |
| CN110674301A (en) | Emotional tendency prediction method, device and system and storage medium | |
| CN110895533B (en) | Form mapping method and device, computer equipment and storage medium | |
| CN116226350A (en) | Document query method, device, equipment and storage medium | |
| CN112560461A (en) | News clue generation method and device, electronic equipment and storage medium | |
| CN115827956A (en) | Data information retrieval method and device, electronic equipment and storage medium | |
| CN112699237B (en) | Label determination method, device and storage medium | |
| CN112560425B (en) | Template generation method and device, electronic equipment and storage medium | |
| CN116484826B (en) | Operation ticket generation method, device, equipment and storage medium | |
| CN117668221A (en) | Vehicle fault handling mode pushing method, device and medium | |
| CN111460088A (en) | Similar text retrieval method, device and system | |
| CN114238634B (en) | Regular expression generation method, application, device, equipment and storage medium | |
| CN117874088B (en) | Data fuzzy matching method, device, equipment and medium | |
| CN115329748B (en) | Log analysis method, device, equipment and storage medium | |
| CN118035445A (en) | Work order classification method and device, electronic equipment and storage medium | |
| CN117807287A (en) | Label fusion method, device, electronic equipment and storage medium | |
| CN115618242A (en) | Repeated text recognition method and device, electronic equipment and storage medium | |
| CN117708758A (en) | Fault cause judging method, device, equipment and medium of fault phenomenon text | |
| CN116542244A (en) | Entity disambiguation method and device for power industry | |
| CN117574168A (en) | Information report generation method and device | |
| CN115391281A (en) | Searching method and device for power document, electronic equipment and storage medium | |
| CN117670554A (en) | Method, device, electronic equipment and storage medium for determining data asset tag | |
| CN116881280A (en) | Optimized database statement determination method, device, equipment and storage medium | |
| CN116308455A (en) | Method and device for identifying hub area in trade network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |