Keyword extraction method and device for rail transit maintenance work order

By calculating the word frequency and reverse file frequency in the maintenance work order corpus, combined with the initial weight, the keywords of rail transit maintenance work order are extracted, and the problems of data diversity and low utilization efficiency are solved, and faster fault cause positioning and maintenance efficiency are achieved. promote.

CN112214990BActive Publication Date: 2024-11-22TRAFFIC CONTROL TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202011015127.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-09-24
Publication Date
2024-11-22
Estimated Expiration
2040-09-24

AI Technical Summary

Technical Problem

The data of rail transit maintenance work orders are large and diverse, and they are not effectively utilized, resulting in low efficiency in determining the cause of failure.

Method used

By obtaining the maintenance work order corpus, calculating the word frequency and reverse file frequency, combining the initial weight and tuning parameters, determining the actual weight, and then extracting the rail transit maintenance work order keywords to achieve effective data classification and problem positioning.

Benefits of technology

It improves the efficiency of rail transit maintenance work order data, can determine the cause of the failure more quickly, and improves the efficiency of the maintenance process.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112214990B_ABST
    Figure CN112214990B_ABST
Patent Text Reader

Abstract

The embodiment of the present invention provides a method and device for extracting keywords from a rail transit maintenance work order. First, a maintenance work order corpus is obtained, and the product of the word frequency of each word in the corpus and the reverse file frequency is calculated; then, based on the initial weight of each word in the maintenance work order corpus and the product of the word frequency of each word and the reverse file frequency, the actual weight of each word in the maintenance work order corpus is determined; finally, according to the actual weight of each word in the maintenance work order corpus, the keywords of the rail transit maintenance work order are determined. The rail transit maintenance work order data is fully utilized, and the keywords of the rail transit maintenance work order are determined by processing the rail transit maintenance work order data, and then the rail transit maintenance work order can be classified by keywords, so that problems in the rail transit maintenance work order can be determined more quickly and the causes of the problems can be found.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of rail transit technology, and in particular to a method and device for extracting keywords from a rail transit maintenance work order. Background Art

[0002] Rail transit technology is developing rapidly. From a maintenance perspective, operation and maintenance are important links in the safe operation of subway systems. In the process of rail transit maintenance, manual work order recording is usually adopted. In the long-term subway operation and maintenance process, the number of work orders accumulated has increased. The work order data records the cause of the fault, the description of the fault, and the solution to the fault, which contains a large amount of information.

[0003] The data provided by rail equipment conforms to the characteristics of big data. The data is diverse, and different rail equipment generates a wide variety of data with complex content at different times in their life cycle. The data formats are diverse, including traditional paper data such as manual line inspection logs, track inspection instrument format data, vehicle on-board controller (VOBC), driving record data, etc. The data structure is diverse, including structured data, semi-structured data, unstructured data, etc. The data scale is huge. Taking the inspection data of the track inspection vehicle as an example, the main line is inspected twice a month, and the corresponding waveform data and over-limit data are stored in the database. The data volume in one year can reach 9TB.

[0004] However, currently work order data is not effectively utilized. If such a large amount of data is not effectively utilized, it will be of no help in determining the cause of the failure. Summary of the invention

[0005] The embodiment of the present invention provides a method and device for extracting keywords from a rail transit maintenance work order, so as to solve the defects existing in the prior art.

[0006] An embodiment of the present invention provides a method for extracting keywords from a rail transit maintenance work order, comprising:

[0007] Obtaining a maintenance work order corpus, calculating the word frequency of each word in the corpus and the reverse file frequency of each word, and determining the product of the word frequency and the reverse file frequency of each word; the maintenance work order corpus is a set of words appearing in a specified number of rail transit maintenance work orders;

[0008] Determining the actual weight of each word in the maintenance work order corpus based on the initial weight of each word in the maintenance work order corpus and the product of the word frequency of each word and the reverse file frequency; the initial weight is determined based on the importance of each word in the maintenance work order corpus;

[0009] Based on the actual weight of each word in the maintenance work order corpus, rail transit maintenance work order keywords are determined.

[0010] The method for extracting keywords from a rail transit maintenance work order according to an embodiment of the present invention further includes:

[0011] Receiving a keyword to be queried, and determining all candidate rail transit maintenance work orders containing the keyword of the rail transit maintenance work order to be queried;

[0012] The weight of each candidate rail transit maintenance work order is calculated, and the candidate rail transit maintenance work order with a higher weight is used as the rail transit maintenance work order corresponding to the keyword to be queried.

[0013] According to a rail transit maintenance work order keyword extraction method according to an embodiment of the present invention, the actual weight of each word in the maintenance work order corpus is determined based on the initial weight of each word in the maintenance work order corpus and the product of the word frequency of each word and the reverse file frequency, specifically including:

[0014] The actual weight is determined based on the initial weight and the product of the word frequency of each word and the reverse document frequency in combination with the tuning parameters.

[0015] According to a rail transit maintenance work order keyword extraction method according to an embodiment of the present invention, the actual weight is determined based on the initial weight and the product of the word frequency of each word and the reverse file frequency, combined with the tuning parameter, and specifically includes:

[0016] Based on the initial weights, the product of the frequency of each word and the inverse document frequency, and the tuning parameters, an intermediate weight is determined;

[0017] Based on the maximum and minimum normalization method, the intermediate weights are normalized to obtain the actual weights.

[0018] According to a method for extracting keywords from a rail transit maintenance work order according to an embodiment of the present invention, the intermediate weight is determined based on the initial weight and the product of the word frequency of each word and the reverse file frequency, combined with a tuning parameter, and specifically includes:

[0019] The intermediate weight is determined by the following formula:

[0020]

[0021] Among them, q i is the intermediate weight of the i-th word, TF is the word frequency of the i-th word, IDF is the inverse document frequency of the i-th word, k is the tuning parameter, is the initial weight of the i-th word.

[0022] According to a rail transit maintenance work order keyword extraction method according to an embodiment of the present invention, the intermediate weights are normalized based on the maximum and minimum normalization method to obtain the actual weights, specifically including:

[0023] The actual weight is determined by the following formula:

[0024]

[0025] Among them, q i-final is the actual weight of the i-th word, q i is the intermediate weight of the i-th word, q min is the minimum value among all intermediate weights, q max is the maximum value among all intermediate weights.

[0026] According to a method for extracting keywords from a rail transit maintenance work order according to an embodiment of the present invention, obtaining a maintenance work order corpus specifically includes:

[0027] For each rail transit maintenance work order of the specified number of rail transit maintenance work orders, the rail transit maintenance work order is processed based on a sentence cleaning method and a stuttering word segmentation method to obtain the maintenance work order corpus.

[0028] The embodiment of the present invention also provides a rail transit maintenance work order keyword extraction device, comprising: a maintenance work order corpus acquisition module, an actual weight determination module and a keyword determination module.

[0029] The maintenance work order corpus acquisition module is used to acquire the maintenance work order corpus, calculate the word frequency of each word in the corpus and the reverse file frequency of each word, and determine the product of the word frequency of each word and the reverse file frequency; the maintenance work order corpus is a set of words appearing in a specified number of rail transit maintenance work orders;

[0030] The actual weight determination module is used to determine the actual weight of each word in the maintenance work order corpus based on the initial weight of each word in the maintenance work order corpus and the product of the word frequency of each word and the reverse file frequency; the initial weight is determined based on the importance of each word in the maintenance work order corpus;

[0031] The keyword determination module is used to determine the rail transit maintenance work order keywords based on the actual weight of each word in the maintenance work order corpus.

[0032] An embodiment of the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of any of the above-mentioned methods for extracting keywords from rail transit maintenance work orders are implemented.

[0033] An embodiment of the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of any of the above-described methods for extracting keywords from rail transit maintenance work orders are implemented.

[0034] The rail transit maintenance work order keyword extraction method and device provided by the embodiment of the present invention first obtains the maintenance work order corpus, and calculates the product of the word frequency of each word in the corpus and the reverse file frequency; then, based on the initial weight of each word in the maintenance work order corpus and the product of the word frequency of each word and the reverse file frequency, the actual weight of each word in the maintenance work order corpus is determined; finally, according to the actual weight of each word in the maintenance work order corpus, the rail transit maintenance work order keywords are determined. The rail transit maintenance work order data is fully utilized, and the rail transit maintenance work order keywords are determined by processing the rail transit maintenance work order data, and then the rail transit maintenance work order can be classified by keywords, so that the problems in the rail transit maintenance work order can be determined more quickly and the causes of the problems can be found. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0036] Figure 1 It is a flow chart of a method for extracting keywords from a rail transit maintenance work order provided by an embodiment of the present invention;

[0037] Figure 2 It is a structural schematic diagram of a rail transit maintenance work order keyword extraction device provided by an embodiment of the present invention;

[0038] Figure 3 It is a structural schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] Figure 1FIG. 1 is a flow chart of a method for extracting keywords from a rail transit maintenance work order provided in an embodiment of the present invention. Figure 1 As shown, the method includes:

[0041] S1, obtaining a maintenance work order corpus, calculating the word frequency of each word in the corpus and the reverse file frequency of each word, and determining the product of the word frequency and the reverse file frequency of each word; the maintenance work order corpus is a set of words appearing in a specified number of rail transit maintenance work orders;

[0042] S2, determining the actual weight of each word in the maintenance work order corpus based on the initial weight of each word in the maintenance work order corpus and the product of the word frequency of each word and the reverse file frequency; the initial weight is determined based on the importance of each word in the maintenance work order corpus;

[0043] S3, determining rail transit maintenance work order keywords based on the actual weight of each word in the maintenance work order corpus.

[0044] Specifically, the rail transit maintenance work order keyword extraction method provided in the embodiment of the present invention aims to determine the rail transit maintenance work order keywords through big data.

[0045] First, execute step S1. Obtain a maintenance work order corpus, which contains all the words that appear in a specified number of rail transit maintenance work orders. The specified number can be set as needed, and in the embodiment of the present invention, it can be considered to be a large number. The words can be single words, or they can be new words formed by the fusion of multiple adjacent words with similar word frequencies, such as words such as "Xizhimen", "subway station", "platform", and new words such as "Xizhimen subway station platform".

[0046] Calculate the term frequency (TF) and inverse document frequency (IDF) of each word in the maintenance work order corpus. Among them, term frequency refers to the frequency of the words existing in the maintenance work order corpus appearing in a certain rail transit maintenance work order. Taking into account the characteristics of rail transit words, the present invention allows the words in the corpus to be fused before and after to form new words in the corpus. For example, "Xizhimen subway station platform" will be divided into three words "Xizhimen", "subway station" and "platform" in word segmentation. However, the single appearance of these three words cannot effectively express the location of the rail transit subway station. Using the proposed fusion method, taking the word "subway station" as an example, the words before and after the "subway station" can be fused together to form a new word "Xizhimen subway station platform", which increases the amount of information represented by the word. The term frequency TF of the i-th word in the maintenance work order corpus i The specific calculation method is:

[0047]

[0048] For example, the term "Xizhimen Subway Station" appears 3 times in a rail transit maintenance work order, which has a total of 1,000 words. Then the term frequency of "Xizhimen Subway Station" is 0.003.

[0049] In the process of calculating word frequency, it can be seen that some common words are not very useful for extracting topic keywords, but they may have a large word frequency, and some words with less frequent occurrence can express the theme of the article. This deviates from the main purpose of keyword extraction, that is, the stronger the words in a rail transit word library can reflect the theme of the work order, the greater the weight, and vice versa. If a large number of work orders are formed into a work order library, some words may only appear in a few articles, but such words can effectively reflect the theme of the article, and these words should be assigned a larger weight. In view of this, the reverse file frequency of each word in the rail transit work order library should be calculated.

[0050] Inverse Document Frequency IDF i The calculation method is:

[0051]

[0052] If the number of rail transit maintenance work orders containing a certain word is greater, the reverse document frequency of the word is smaller; if the number of rail transit maintenance work orders containing a certain word is smaller, the reverse document frequency of the word is greater. The purpose of +1 in the denominator is to prevent the phenomenon that a certain word in the maintenance work order corpus does not appear in a certain rail transit maintenance work order, making the reverse document frequency value of the word infinitely small.

[0053] After calculating the word frequency and reverse document frequency of all words in the rail transit work order database, it is necessary to calculate the word frequency-reverse document frequency (TF-IDF) of each word. TF-IDF combines the frequency of words in rail transit maintenance work orders and the reverse document frequency of work orders in the entire rail transit work order database, which can filter out common words and retain important words as keywords.

[0054] TF-IDF of the i-th word i The calculation method is:

[0055] TF-IDF i =TF*IDF;

[0056] From the above formula, we can see that when calculating the TF-IDF of the i-th word iIn the process, no relevant weighting is performed on any word, that is, all words are considered to be of equal importance except for the calculation of word frequency and inverse document frequency. However, considering the scenario of keyword extraction of rail transit maintenance work orders, it can be seen that different words in rail transit maintenance work orders have different importance. For example, the word "switch machine" is obviously more targeted than "subway station", and auxiliary information should be added, and higher weights should be given to various proper nouns to make the extracted keywords more reasonable. Therefore, it is necessary to improve the calculation method of TF-IDF, and combine the actual application scenarios to appropriately reshape and transform the above formula to make it more suitable for the application scenario of rail transit maintenance work orders.

[0057] To this end, in the embodiment of the present invention, when executing step S2, the initial weight of each word in the maintenance work order corpus is introduced, and the initial weight is determined based on the importance of each word in the maintenance work order corpus. The initial weight of each word is combined with TF-IDF to obtain the actual weight of each word in the maintenance work order corpus.

[0058] Finally, step S3 is performed to determine the rail transit maintenance work order keywords according to the actual weight of each word in the maintenance work order corpus. Specifically, the words in the maintenance work order corpus can be sorted according to the actual weight of each word, and several words with larger actual weights are selected as rail transit maintenance work order keywords.

[0059] The rail transit maintenance work order keyword extraction method provided in the embodiment of the present invention first obtains the maintenance work order corpus, and calculates the product of the word frequency of each word in the corpus and the reverse file frequency; then, based on the initial weight of each word in the maintenance work order corpus, and the product of the word frequency of each word and the reverse file frequency, the actual weight of each word in the maintenance work order corpus is determined; finally, according to the actual weight of each word in the maintenance work order corpus, the rail transit maintenance work order keyword is determined. The rail transit maintenance work order data is fully utilized, and the rail transit maintenance work order keywords are determined by processing the rail transit maintenance work order data, and then the rail transit maintenance work order can be classified by keywords, so that the problems in the rail transit maintenance work order can be determined more quickly and the causes of the problems can be found.

[0060] On the basis of the above-mentioned embodiment, the rail transit maintenance work order keyword extraction method provided in the embodiment of the present invention further includes:

[0061] Receiving a keyword to be queried, and determining all candidate rail transit maintenance work orders containing the keyword of the rail transit maintenance work order to be queried;

[0062] The weight of each candidate rail transit maintenance work order is calculated, and the candidate rail transit maintenance work order with a higher weight is used as the rail transit maintenance work order corresponding to the keyword to be queried.

[0063] Specifically, the rail transit maintenance work order keyword extraction method provided in the embodiment of the present invention can determine the matching work order based on the keywords input by the maintenance personnel on the basis of determining the rail transit maintenance work order keywords.

[0064] First, a keyword to be queried is received, where the keyword to be queried is a keyword input by a maintenance person, and a specified number of rail transit maintenance work orders containing the keyword to be queried are determined and used as candidate rail transit maintenance work orders.

[0065] Then, the weight of each candidate rail transit maintenance work order is calculated. Specifically, the word frequency of the keyword to be queried in each candidate rail transit maintenance work order can be determined, and the word frequency is used as the weight of the candidate rail transit maintenance work order.

[0066] Finally, the alternative rail transit maintenance work order with a higher weight is used as the rail transit maintenance work order corresponding to the keyword to be queried, and the alternative rail transit maintenance work order can be pushed to the maintenance personnel to achieve keyword query matching.

[0067] In the embodiment of the present invention, by determining the rail transit maintenance work order corresponding to the keyword to be queried, it is possible to help maintenance personnel quickly determine the cause of the problem.

[0068] On the basis of the above-mentioned embodiment, the rail transit maintenance work order keyword extraction method provided in the embodiment of the present invention, based on the initial weight of each word in the maintenance work order corpus, and the product of the word frequency of each word and the reverse file frequency, determines the actual weight of each word in the maintenance work order corpus, specifically including:

[0069] The actual weight is determined based on the initial weight and the product of the word frequency of each word and the reverse document frequency in combination with the tuning parameters.

[0070] Specifically, in the embodiment of the present invention, when determining the actual weight of each word in the maintenance work order corpus, the intermediate weight can be determined by the following formula based on the initial weight and the product of the word frequency of each word and the reverse file frequency, combined with the tuning parameters:

[0071]

[0072] Among them, q i is the intermediate weight of the i-th word, TF is the word frequency of the i-th word, IDF is the inverse document frequency of the i-th word, k is the tuning parameter, is the initial weight of the i-th word.

[0073] Then, the intermediate weight is used as the actual weight of the i-th word.

[0074] On the basis of the above-mentioned embodiments, in the rail transit maintenance work order keyword extraction method provided in the embodiments of the present invention, the initial weight can be determined by the expert's weight scoring of the importance of each word, that is, the expert scoring method is used to determine the initial weight of each word in the maintenance work order corpus. The weight scores of all words form a weight matrix W, and the specific form of the weight matrix W is as follows:

[0075]

[0076] Where m is the number of experts, n is the number of all words in the maintenance work order corpus, and w ij It represents the weight score of the jth expert on the ith word. In this way, the weight of keywords in the field of rail transit, especially in the maintenance direction, can be effectively improved.

[0077] The initial weight of the jth word in the maintenance work order corpus can be expressed as:

[0078]

[0079] In the embodiment of the present invention, since the expert's weighted scoring is mainly aimed at highlighting rail transit-specific vocabulary, the rail transit-specific vocabulary in the maintenance work order corpus has a higher weighted score, while non-rail transit-specific vocabulary has a lower weighted score.

[0080] On the basis of the above-mentioned embodiment, the rail transit maintenance work order keyword extraction method provided in the embodiment of the present invention, based on the initial weight and the product of the word frequency of each word and the reverse file frequency, combined with the tuning parameter, determines the actual weight, specifically including:

[0081] Based on the initial weights, the product of the frequency of each word and the inverse document frequency, and the tuning parameters, an intermediate weight is determined;

[0082] Based on the maximum and minimum normalization method, the intermediate weights are normalized to obtain the actual weights.

[0083] Specifically, in the embodiment of the present invention, when determining the actual weight, the intermediate weight can be normalized based on the determination of the intermediate weight and combined with the maximum and minimum normalization method to obtain the final actual weight.

[0084] The actual weight can be determined by the following formula:

[0085]

[0086] Among them, q i-fin is the actual weight of the i-th word, q iis the intermediate weight of the i-th word, q min is the minimum value among all intermediate weights, q max is the maximum value among all intermediate weights.

[0087] Based on the above embodiments, in the embodiments of the present invention, for the method for extracting keywords from rail transit maintenance work orders, the obtaining of the maintenance work order corpus specifically includes:

[0088] For each rail transit maintenance work order among the specified number of rail transit maintenance work orders, based on the sentence cleaning method and the Jieba word segmentation method, process the rail transit maintenance work order to obtain the maintenance work order corpus.

[0089] Specifically, when obtaining the maintenance work order corpus, each rail transit maintenance work order among the specified number of rail transit maintenance work orders can be used as a sample. The sentence cleaning method is adopted to clean special characters such as full stops, colons, commas, etc. At the same time, common words such as "of", "is", "have", etc. are removed. Then, the Jieba word segmentation method is called to convert the rail transit maintenance work order data from unstructured text data into standardized single words for storage, and the maintenance work order corpus is constructed.

[0090] Figure 2 is a schematic structural diagram of a device for extracting keywords from rail transit maintenance work orders provided in an embodiment of the present invention, as Figure 2 shown. The device includes: a maintenance work order corpus obtaining module 21, an actual weight determining module 22, and a keyword determining module 23. Among them,

[0091] The maintenance work order corpus obtaining module 21 is used to obtain the maintenance work order corpus, calculate the word frequency of each word in the corpus and the inverse document frequency of each word, and determine the product of the word frequency and the inverse document frequency of each word; the maintenance work order corpus is a set containing the words that appear in the specified number of rail transit maintenance work orders;

[0092] The actual weight determining module 22 is used to determine the actual weight of each word in the maintenance work order corpus based on the initial weight of each word in the maintenance work order corpus and the product of the word frequency and the inverse document frequency of each word; the initial weight is determined based on the importance of each word in the maintenance work order corpus;

[0093] The keyword determining module 23 is used to determine the rail transit maintenance work order keywords based on the actual weight of each word in the maintenance work order corpus.

[0094] Specifically, the functions of each module in the rail transit maintenance work order keyword extraction device provided in the embodiment of the present invention correspond one-to-one to the operating procedures of each step in the above-mentioned method embodiment, and the effects achieved are also consistent. Please refer to the above-mentioned embodiment for details, and no further details will be given in the embodiment of the present invention.

[0095] Figure 3 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 3 As shown, the electronic device may include: a processor (processor) 310, a communication interface (Communications Interface) 320, a memory (memory) 330 and a communication bus 340, wherein the processor 310, the communication interface 320, and the memory 330 communicate with each other through the communication bus 340. The processor 310 can call the logic instructions in the memory 330 to execute the rail transit maintenance work order keyword extraction method, including: obtaining a maintenance work order corpus, calculating the word frequency of each word in the corpus and the reverse file frequency of each word, and determining the product of the word frequency of each word and the reverse file frequency; the maintenance work order corpus is a set of words that appear in a specified number of rail transit maintenance work orders; based on the initial weight of each word in the maintenance work order corpus, and the product of the word frequency of each word and the reverse file frequency, determine the actual weight of each word in the maintenance work order corpus; the initial weight is determined based on the importance of each word in the maintenance work order corpus; based on the actual weight of each word in the maintenance work order corpus, determine the rail transit maintenance work order keywords.

[0096] In addition, the logic instructions in the above-mentioned memory 330 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0097] On the other hand, an embodiment of the present invention further provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the rail transit maintenance work order keyword extraction method provided by the above-mentioned method embodiments, including: obtaining a maintenance work order corpus, calculating the word frequency of each word in the corpus and the reverse file frequency of each word, and determining the product of the word frequency of each word and the reverse file frequency; the maintenance work order corpus is a set of words appearing in a specified number of rail transit maintenance work orders; based on the initial weight of each word in the maintenance work order corpus, and the product of the word frequency of each word and the reverse file frequency, determining the actual weight of each word in the maintenance work order corpus; the initial weight is determined based on the importance of each word in the maintenance work order corpus; based on the actual weight of each word in the maintenance work order corpus, determining rail transit maintenance work order keywords.

[0098] On the other hand, an embodiment of the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the rail transit maintenance work order keyword extraction method provided in the above-mentioned embodiments, including: obtaining a maintenance work order corpus, calculating the word frequency of each word in the corpus and the reverse file frequency of each word, and determining the product of the word frequency of each word and the reverse file frequency; the maintenance work order corpus is a set of words appearing in a specified number of rail transit maintenance work orders; based on the initial weight of each word in the maintenance work order corpus, and the product of the word frequency of each word and the reverse file frequency, determining the actual weight of each word in the maintenance work order corpus; the initial weight is determined based on the importance of each word in the maintenance work order corpus; based on the actual weight of each word in the maintenance work order corpus, determining rail transit maintenance work order keywords.

[0099] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0100] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0101] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for extracting keywords from rail transit maintenance work orders, characterized in that: include: Obtaining a maintenance work order corpus, calculating the word frequency of each word in the corpus and the reverse file frequency of each word, and determining the product of the word frequency of each word and the reverse file frequency; The maintenance work order corpus is a set of words appearing in a specified number of rail transit maintenance work orders; Determining the actual weight of each word in the maintenance work order corpus based on the initial weight of each word in the maintenance work order corpus and the product of the word frequency of each word and the reverse file frequency; The initial weight is determined based on the importance of each word in the maintenance work order corpus; Determining rail transit maintenance work order keywords based on the actual weight of each word in the maintenance work order corpus; The determining of the actual weight of each word in the maintenance work order corpus based on the initial weight of each word in the maintenance work order corpus and the product of the word frequency of each word and the reverse file frequency specifically includes: Based on the initial weight and the product of the word frequency of each word and the inverse document frequency, combined with the tuning parameters, the actual weight is determined; The actual weight is determined based on the initial weight, the product of the word frequency of each word and the reverse file frequency, and the tuning parameters, and specifically includes: Based on the initial weights, the product of the frequency of each word and the inverse document frequency, and the tuning parameters, an intermediate weight is determined; Based on the maximum and minimum normalization method, the intermediate weights are normalized to obtain the actual weights; The method of determining the intermediate weight based on the initial weight, the product of the word frequency of each word and the reverse file frequency, and the tuning parameters specifically includes: The intermediate weight is determined by the following formula: in, is the intermediate weight of the i-th word, TF is the word frequency of the i-th word, IDF is the inverse document frequency of the i-th word, k is the tuning parameter, is the initial weight of the i-th word.

2. The method for extracting keywords from rail transit maintenance work orders according to claim 1, characterized in that: Also includes: Receiving a keyword to be queried, and determining all candidate rail transit maintenance work orders containing the keyword of the rail transit maintenance work order to be queried; The weight of each candidate rail transit maintenance work order is calculated, and the candidate rail transit maintenance work order with a higher weight is used as the rail transit maintenance work order corresponding to the keyword to be queried.

3. The method for extracting keywords from rail transit maintenance work orders according to claim 1, characterized in that: The step of normalizing the intermediate weights based on the maximum-minimum normalization method to obtain the actual weights specifically includes: The actual weight is determined by the following formula: in, is the actual weight of the i-th word, is the intermediate weight of the i-th word, is the minimum value among all intermediate weights, is the maximum value among all intermediate weights.

4. The method for extracting keywords from a rail transit maintenance work order according to any one of claims 1 to 3, characterized in that: The obtaining of the maintenance work order corpus specifically includes: For each rail transit maintenance work order of the specified number of rail transit maintenance work orders, the rail transit maintenance work order is processed based on a sentence cleaning method and a stuttering word segmentation method to obtain the maintenance work order corpus.

5. A rail transit maintenance work order keyword extraction device, characterized in that: include: A maintenance work order corpus acquisition module, used to acquire a maintenance work order corpus, calculate the word frequency of each word in the corpus and the reverse file frequency of each word, and determine the product of the word frequency of each word and the reverse file frequency; The maintenance work order corpus is a set of words appearing in a specified number of rail transit maintenance work orders; An actual weight determination module, configured to determine an actual weight of each word in the maintenance work order corpus based on an initial weight of each word in the maintenance work order corpus and a product of a word frequency of each word and a reverse file frequency; The initial weight is determined based on the importance of each word in the maintenance work order corpus; A keyword determination module, used to determine rail transit maintenance work order keywords based on the actual weight of each word in the maintenance work order corpus; The actual weight determination module is specifically used for: Based on the initial weight and the product of the word frequency of each word and the inverse document frequency, combined with the tuning parameters, the actual weight is determined; The actual weight is determined based on the initial weight, the product of the word frequency of each word and the reverse file frequency, and the tuning parameters, and specifically includes: Based on the initial weights, the product of the frequency of each word and the inverse document frequency, and the tuning parameters, an intermediate weight is determined; Based on the maximum and minimum normalization method, the intermediate weights are normalized to obtain the actual weights; The method of determining the intermediate weight based on the initial weight, the product of the word frequency of each word and the reverse file frequency, and the tuning parameters specifically includes: The intermediate weight is determined by the following formula: in, is the intermediate weight of the i-th word, TF is the word frequency of the i-th word, IDF is the inverse document frequency of the i-th word, k is the tuning parameter, is the initial weight of the i-th word.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the rail transit maintenance work order keyword extraction method as described in any one of claims 1 to 4 are implemented.

7. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for extracting keywords from a rail transit maintenance work order as described in any one of claims 1 to 4 are implemented.