Intelligent answering recognition device and method applied to railway production dispatching incoming call

Abstract

The invention relates to the technical field of answering recognition, in particular to an intelligent answering recognition device and method applied to a railway production dispatching incoming call, and the method specifically comprises the steps: obtaining the voice information of the railway production dispatching incoming call at the current moment and the information record of each historical railway production dispatching incoming call; on the basis of the occurrence condition of each keyword in the current scheduling task in each historical scheduling incoming call information record and the disorder condition of the occurrence frequency of all keywords in the current scheduling task, the complexity of the current scheduling incoming call voice information is obtained; according to the method, the number of events selected during retrieval is determined based on the complexity of current scheduling incoming call voice information, so that a more adaptive reference historical scheduling record is determined for a current scheduling task for reference of a dispatcher; the problem that a final retrieval result cannot be better referred to a dispatcher due to the fact that a traditional detection fine standard cannot be determined is solved, and the dispatching response efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of call answering and recognition technology, and in particular to an intelligent call answering and recognition device and method for railway production dispatching. Background Art

[0002] Intelligent call recognition devices and methods for railway production dispatching are a key aspect of the intelligent and automated development of the railway industry. With the growing demand for railway transportation and the increasing complexity of operational management, traditional dispatching methods are no longer able to meet the needs for efficient and accurate dispatching. Therefore, the introduction of intelligent call recognition technology has become an inevitable trend in the industry's development. Currently, the railway industry is actively exploring and applying various intelligent technologies to improve dispatching efficiency and safety. For example, voice recognition technology converts dispatchers' voice commands into text messages, which are then analyzed and processed using natural language processing technology, enabling real-time monitoring of train operating status and automated execution of dispatching instructions.

[0003] Currently, dispatchers typically respond to incoming calls from railway production dispatchers according to established rules and procedures. This can hinder the efficiency of railway production dispatchers, as dispatchers vary in their proficiency and experience with the work they do. Furthermore, dispatchers typically review similar dispatch records to identify helpful solutions or recommendations. However, there's currently no way to determine appropriate screening criteria based on the incoming call and provide dispatchers with appropriate reference solutions, hindering the efficiency of production dispatchers. Summary of the Invention

[0004] In order to solve the above technical problems, the purpose of this application is to provide an intelligent answering and identification device and method for railway production dispatching. The technical solutions adopted are as follows:

[0005] In a first aspect, an embodiment of the present application provides a method for intelligently answering and identifying incoming calls for railway production scheduling, the method comprising the following steps:

[0006] Obtain the voice information of the current railway production dispatch call and the information record of each railway production dispatch call in the past, record them as the current dispatch call voice information and the historical dispatch call information record, and obtain keywords for each;

[0007] Record the text content of the current scheduled incoming call voice information as the current scheduled task; determine the event occurrence frequency of the current scheduled task based on the occurrence of each keyword in the current scheduled task;

[0008] Determine the degree of event irrelevance of the current scheduling task based on the occurrence of each keyword in the history of each scheduling call information record; construct the event low-frequency consistency of the current scheduling task based on the disorder of the occurrence frequency of all keywords in the current scheduling task; construct the complexity of the current scheduling task based on the event occurrence frequency, event irrelevance and event low-frequency consistency of the current scheduling task;

[0009] The preset length of time each day is considered as a cycle; the sensitivity of the peak period of each cycle of the day is calculated based on the number of railway production scheduling calls in each cycle of the day; the number of subsequent event retrievals for the current scheduling task is determined based on the complexity of the current scheduling task, the number of keywords, and the sensitivity during the peak period;

[0010] Based on the number of subsequent event retrievals, keywords for subsequent retrieval in the current scheduling task are determined.

[0011] In one embodiment, the process of obtaining keywords is as follows:

[0012] The railway production scheduling knowledge base is used as the input of the keyword extraction algorithm, and the output is the keywords in the scheduling knowledge base; for any word in the scheduling problem of the current scheduling call voice information and the historical scheduling call information records, if the any word is the same as any keyword in the scheduling knowledge base, then the any word is used as the keyword in the scheduling problem corresponding to the current scheduling call voice information or the historical scheduling call information records.

[0013] In one embodiment, the event occurrence frequency is: the average occurrence frequency of all keywords in the current scheduling task.

[0014] In one embodiment, the process of obtaining the degree of irrelevance of events is as follows:

[0015] For keyword i in the current scheduling task, obtain the number of times each keyword and keyword i appear together in all historical scheduling call information records, and record it as the first absolute number; the maximum value of all the first absolute numbers of keyword i is taken as the absolute number of occurrences of keyword i;

[0016] The process of obtaining the relative number of occurrences of each keyword in the current scheduling task is:

[0017] P1, randomly obtains the first keyword from the current scheduled task and records it as the first keyword;

[0018] P2, classify the first keyword into the determined keywords;

[0019] P3, all the remaining keywords except the determined keywords in the current scheduling task will be obtained as undetermined keywords;

[0020] P4, obtaining the number of co-occurrences of each undetermined keyword and the first keyword in all historical dispatch call information records, recording it as the first relative number; obtaining the maximum value of all the first relative numbers of the first keyword as the relative number of occurrences of the first keyword; recording the undetermined keyword corresponding to the maximum value as the second keyword;

[0021] P5: Classify the second keyword into the determined keywords and repeat steps P3-P5 until the corresponding relative number of occurrences of each keyword in the current scheduling task is obtained;

[0022] The degree of irrelevance of the events of the current scheduled task is recorded as W, and the expression of W is: Where I represents the number of keywords in the current scheduling task, w′ i is the relative number of occurrences of the i-th keyword in the current scheduling task, w i The absolute number of occurrences of the i-th keyword in the current scheduled task.

[0023] In one embodiment, the expression of the event low-frequency consistency is:

[0024] E=1-norm(σ q ), where E represents the low-frequency consistency of the event of the current scheduling task, norm() is the normalization function, σ q is the standard deviation of the frequency of occurrence of all keywords in the current scheduling task.

[0025] In one embodiment, the complexity of the current scheduled task is expressed as:

[0026] R = Q × W + (1-Q) × E, where R represents the complexity of the current scheduling task, Q represents the frequency of events in the current scheduling task, W represents the degree of irrelevance of the events in the current scheduling task, and E represents the low-frequency consistency of the events in the current scheduling task.

[0027] In one embodiment, the process of obtaining the sensitivity during the peak period of each cycle in a day is as follows:

[0028] Calculate the ratio of the number of railway production dispatch calls in each cycle of each day to the total number of railway production dispatch calls on that day as the peak proportion of railway production dispatch calls in each cycle of each day; record the sensitivity during the peak period of the kth cycle in a day as U k , U k The expression is:

[0029] In the formula, norm() is the normalization function, P represents the preset number of days, and x p,kIt represents the peak proportion of railway production scheduling calls in the k-th cycle on the p-th day.

[0030] In one embodiment, the expression for the number of subsequent event retrievals for the current scheduled task is:

[0031] Where J represents the number of subsequent event retrievals for the current scheduling task, I represents the number of keywords in the current scheduling task, R represents the complexity of the current scheduling task, and U′ represents the sensitivity during the peak period of the cycle at the current moment. Indicates the rounding symbol, norm() is the normalization function.

[0032] In one embodiment, the process of determining the keywords for subsequent retrieval in the current scheduling task is as follows:

[0033] The number of subsequent event retrievals for the current scheduling task is recorded as the first number, all keywords in the current scheduling task are arranged from high to low according to the event occurrence frequency of the keywords to form a keyword sequence, and the first number of keywords in the keyword sequence are used as keywords for subsequent retrieval.

[0034] In the second aspect, an embodiment of the present application also provides an intelligent answering and identification device for railway production scheduling calls, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the steps of any one of the above methods are implemented.

[0035] The embodiments of the present application have at least the following beneficial effects:

[0036] The present application obtains the voice information of the railway production dispatch call at the current moment and the information record of each historical railway production dispatch call, and obtains the complexity of the current dispatch call voice information based on the appearance of each keyword in the current dispatch task in the historical dispatch call information record, as well as the confusion of the appearance frequency of all keywords of the current dispatch task. At the same time, the complexity of the dispatch task is taken into account under the two situations of high and low overall occurrence frequency of events, thereby improving the rationality of setting the number of keyword retrievals when the complexity of events is adaptively set to different levels of complexity. The application determines the number of events selected for retrieval based on the complexity of the current dispatch call voice information, thereby determining a more suitable reference historical dispatch record for the current dispatch task for the dispatcher to refer to, thereby avoiding the problem that the final retrieval results cannot provide a better reference for the dispatcher due to the inability to determine the fine standards for detection, thereby improving the dispatch response efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0038] Figure 1 A flowchart of the steps of a method for intelligently answering and identifying incoming calls for railway production scheduling provided in one embodiment of the present application;

[0039] Figure 2 Schematic diagram of the process of determining keywords for subsequent retrieval. DETAILED DESCRIPTION

[0040] To further illustrate the technical means and effects adopted by this application to achieve the intended invention objectives, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features, and effects of the device and method for intelligently answering and identifying incoming calls for railway production scheduling proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable form.

[0041] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0042] The specific scheme of the device and method for intelligent answering and identifying incoming calls for railway production dispatching provided by this application is described in detail below with reference to the accompanying drawings.

[0043] See also Figure 1 , which shows a flowchart of a method for intelligently answering and identifying incoming calls for railway production scheduling provided by an embodiment of the present application, the method comprising the following steps:

[0044] Step S1, obtaining the voice information of the current railway production dispatch call and the information record of each railway production dispatch call in history, recording them as the current dispatch call voice information and the historical dispatch call information record, and obtaining keywords respectively.

[0045] If the dispatcher receives a call from the railway production dispatcher at the current moment, they will obtain the voice information from the call, record it as the current dispatcher call voice information, and use iFlytek's speech recognition technology to convert it into text for subsequent analysis and processing. iFlytek's speech recognition technology is well known, so the specific process will not be repeated here.

[0046] It should be noted that for speech recognition, this application only provides a speech recognition method. There are many existing speech recognition methods, and implementers can also use other speech recognition technologies to convert speech into text. This application does not make specific restrictions.

[0047] At the same time, when the railway production dispatcher calls, the dispatcher will record the content of the call. Therefore, the information of each railway production dispatcher call in the historical time is obtained according to the dispatcher's call record, and recorded as the historical dispatcher call information record. The call information of each record includes the scheduling problem and the corresponding solution.

[0048] The text content corresponding to the current dispatch call voice information and the dispatch problem in each historical dispatch call information record are respectively segmented using the JIEBA word segmentation technology; the railway production dispatch knowledge base is the prior knowledge about railway production dispatch, and the railway production dispatch knowledge base is used as the input of the TF-IDF algorithm to extract keywords, and the output is the keywords contained in the dispatch knowledge base; if each word in the text content of the current dispatch call voice information is the same as any keyword in the dispatch knowledge base, then the word is the keyword in the text content of the current dispatch call voice information; similarly, if each word in the dispatch problem of each historical dispatch call information record is the same as any keyword in the dispatch knowledge base, then the word is the keyword in the dispatch problem of each historical dispatch call information record. Among them, the JIEBA word segmentation technology and the TF-IDF algorithm are both well-known technologies, and the specific process will not be repeated.

[0049] Step S2: Record the text content of the currently scheduled incoming call voice information as the currently scheduled task; and determine the event occurrence frequency of the currently scheduled task based on the occurrence of each keyword in the currently scheduled task.

[0050] For railway production dispatching, employees at various production systems involved in the railway will report repairs or reflect various types of problems through phone calls, such as passenger, freight, passenger, etc. Such complicated types of problems will bring certain dispatching pressure to the dispatcher. In order to improve the dispatcher's work efficiency, the current railway call voice information is matched with appropriate historical dispatch call information records, which provides certain reference and help for the dispatcher, thereby improving the dispatch response speed.

[0051] Since the keywords contained in the task content reported in the dispatch call can determine the type of task that needs to be handled in the dispatch call, such as train operation, cargo loading and unloading, repair and maintenance, etc., each task has its own specific requirements and processes, and the number of keywords in the dispatch call determines the complexity of the dependencies between the dispatch tasks. For example, the delay of a train may affect the operation plan of other trains, and it means that there are more resource constraints. For example, the availability and distribution of resources such as tracks, stations, locomotives, and personnel need to be comprehensively considered to formulate scheduling plans based on these dependencies. The frequency of occurrence of keywords can reflect the historical comparison cases of similar situations. Therefore, the frequency of occurrence of events in the current dispatch call voice message is determined based on the keywords in the text content of the current dispatch call voice message. The method is:

[0052] The text content of the currently scheduled incoming call voice message is recorded as the current scheduled task; the occurrence frequency of each keyword in the current scheduled task is obtained, and the expression of the event occurrence frequency of the current scheduled task is:

[0053] Where Q represents the frequency of events in the current scheduling task, I represents the number of keywords in the current scheduling task, and q i Indicates the frequency of occurrence of the i-th keyword in the current scheduling task.

[0054] Step S3, based on the appearance of each keyword in the history of each scheduled call information record of the current scheduling task, determine the degree of event irrelevance of the current scheduling task; based on the confusion of the appearance frequency of all keywords in the current scheduling task, construct the event low-frequency consistency of the current scheduling task; based on the event appearance frequency, event irrelevance degree and event low-frequency consistency of the current scheduling task, construct the complexity of the current scheduling task.

[0055] A keyword corresponds to an event. In daily railway dispatching work, certain events are usually correlated. That is, certain events usually occur together in dispatching work, and the keywords corresponding to these events usually appear together in dispatching call information. Therefore, when there are events that do not often appear together in the dispatching task, it means that this may correspond to a more complex and difficult dispatching task.

[0056] To analyze the mutual influence between various events, first, obtain the absolute number of occurrences of each keyword in the current scheduling task, specifically:

[0057] Taking keyword i in the current scheduling task as an example, obtain the number of times each keyword and keyword i appear together in all historical scheduling call information records, and record it as the first absolute number; take the maximum value of all the first absolute numbers of keyword i as the absolute number of occurrences of keyword i.

[0058] The co-occurrence means that if keyword i appears in a certain historical dispatch call information record, all other keywords except keyword i in the certain historical dispatch call information record are keywords that co-occur with keyword i;

[0059] The number of co-occurrences means that if keyword i and a certain keyword co-appear in X historical scheduling call information records, the number of co-occurrences of keyword i and the certain keyword is X.

[0060] Then, get the relative number of occurrences of each keyword in the current scheduling task, specifically:

[0061] P1, randomly obtains the first keyword from the current scheduled task and records it as the first keyword;

[0062] P2, classify the first keyword into the determined keywords;

[0063] P3, all the remaining keywords except the determined keywords in the current scheduling task will be obtained as undetermined keywords;

[0064] P4, obtaining the number of co-occurrences of each undetermined keyword and the first keyword in all historical dispatch call information records, recording it as the first relative number; obtaining the maximum value of all the first relative numbers of the first keyword as the relative number of occurrences of the first keyword; recording the undetermined keyword corresponding to the maximum value as the second keyword;

[0065] P5, classify the second keyword into the determined keywords, and repeat steps P3-P5 until each keyword in the current scheduling task obtains the corresponding relative number of occurrences.

[0066] Among them, for the last keyword in the previous scheduling task, if its relative occurrence count cannot be obtained through the above method, the common occurrence count of the last keyword and each of the remaining keywords in the current scheduling task is obtained, and the maximum value is calculated to obtain the relative occurrence count of the last keyword.

[0067] Finally, the expression of the degree of irrelevance of the events of the current scheduled task is:

[0068] Where W represents the degree of irrelevance of the events in the current scheduling task, I represents the number of keywords in the current scheduling task, and w′ i is the relative number of occurrences of the i-th keyword in the current scheduling task, w i The absolute number of occurrences of the i-th keyword in the current scheduled task.

[0069] Furthermore, calculate the low-frequency consistency of events of the current scheduled task:

[0070] In addition to the above analysis of the degree of irrelevance of the time of the current scheduling task, the low-frequency consistency of the current scheduling task should also be considered. This means that when the events contained in the current scheduling task all maintain an extremely low frequency of occurrence, it indicates that the current scheduling task may be facing a relatively rare situation and lacks the support and assistance of historical reference information. By analyzing the low-frequency consistency, we can avoid the situation where the overall low frequency of the current scheduling task is caused by the low frequency of an event. The expression of the low-frequency consistency of the events of the current scheduling task is:

[0071] E=1-norm(σ q ), where E represents the low-frequency consistency of the event of the current scheduling task, norm() is the normalization function, σ q is the standard deviation of the frequency of occurrence of all keywords in the current scheduling task.

[0072] Furthermore, the complexity of the current scheduling task is calculated:

[0073] The complexity of the current scheduling task can be determined through the above process. The frequency of events in the current scheduling task is used as a benchmark. When the frequency of events is high, the complexity of the scheduling task under high frequency should be analyzed from the perspective of the simultaneous occurrence of common related events. When the frequency of events is higher, more attention should be paid to the degree of irrelevance of the events in the current scheduling task. When the frequency of events is low, the complexity of the scheduling task should be analyzed from the perspective of the rarity of events. The lower the frequency of events, the less reference information there is, and the more attention should be paid to the low-frequency consistency of the events in the current scheduling task. Based on this, the complexity of the current scheduling task is constructed, and the expression is:

[0074] R = Q × W + (1-Q) × E, where R represents the complexity of the current scheduling task, Q represents the frequency of events in the current scheduling task, W represents the degree of irrelevance of the events in the current scheduling task, and E represents the low-frequency consistency of the events in the current scheduling task.

[0075] Step S4, taking the preset time length of each day as a cycle; calculating the sensitivity during the peak period of each cycle of the day based on the number of railway production scheduling calls in each cycle of the day; determining the number of subsequent event retrievals for the current scheduling task based on the complexity of the current scheduling task, the number of keywords and the sensitivity during the peak period.

[0076] In railway production scheduling, the degree of progress varies in different time periods, leading to different problems. During peak periods, continuous production operations lead to production docking and coordination issues. While daily production operations may vary, the overall operating pattern remains relatively stable. Therefore, peak periods can be defined by analyzing the number of production scheduling calls received during each recent period.

[0077] First, the analysis is conducted over the past week, with a half-hour cycle length per day. For example, 0.00-0.30 is the first cycle, and 0.31-1.00 is the second cycle. The number of railway production dispatch calls in each cycle of each day is obtained.

[0078] It should be noted that, with respect to the setting of the number of days and the length of the cycle, in other embodiments of the present application, the implementer may set the number of days and the length of the cycle according to actual conditions, and this application does not impose any specific restrictions.

[0079] Furthermore, the ratio of the number of railway production dispatch calls in each cycle of each day to the total number of railway production dispatch calls on that day is calculated as the peak proportion of railway production dispatch calls in each cycle of each day;

[0080] Furthermore, the expression of sensitivity during the peak period of each cycle in a day is:

[0081]

[0082] Where U k represents the sensitivity during the peak period of the kth cycle in a day; norm() is a normalization function; P represents the preset number of days, and the value of P in the embodiment of the present application is 7; x p,k It represents the peak proportion of railway production scheduling calls in the k-th cycle on the p-th day, where the sensitivity during the peak period of the k-th cycle on different days is the same.

[0083] For each cycle of railway production scheduling, the peak proportion of daily production scheduling calls in the recent period is higher, which means that the cycle is in the peak period of railway production scheduling, that is, the sensitivity is higher during the peak period.

[0084] Then, the complexity of the current scheduling task can be determined through the above process, based on which the degree of refinement of railway production scheduling screening can be adjusted. The higher the complexity of the current scheduling task and the peak period, the higher the screening refinement should be, so as to ensure that the historical scheduling records provided to the dispatcher are more adaptable and improve the dispatcher's work efficiency.

[0085] The expression for determining the number of subsequent event retrievals based on the complexity of the current scheduling task is:

[0086] Where J represents the number of subsequent event retrievals for the current scheduling task, I represents the number of keywords in the current scheduling task, R represents the complexity of the current scheduling task, and U′ represents the sensitivity during the peak period of the cycle at the current moment. Indicates the rounding symbol, norm() is the normalization function.

[0087] This determines the number of subsequent retrieval events in the current scheduling task, thereby meeting different levels of screening sophistication.

[0088] Step S5: determining keywords for subsequent retrieval in the current scheduling task based on the number of subsequent event retrievals.

[0089] Arrange all keywords in the current scheduling task according to their event occurrence frequency from high to low to form a keyword sequence, and use the first J keywords in the keyword sequence as the keywords for subsequent retrieval.

[0090] Filter by the frequency of events in the current dispatch task. Events with higher frequency are more likely to be used for search, as they typically correspond to more historical references, making them more helpful for dispatchers to reference and compare. Furthermore, by varying the level of filtering sophistication, we provide dispatchers with different levels of similar dispatch tasks for reference, thus achieving a balance between the quantity and accuracy of historical dispatch records provided to dispatchers.

[0091] The schematic diagram of the process of determining keywords for subsequent retrieval is as follows Figure 2 shown.

[0092] Based on the same inventive concept as the above method, an embodiment of the present application also provides an intelligent answering and identification device for railway production scheduling calls, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-mentioned methods for intelligent answering and identification of railway production scheduling calls.

[0093] In summary, the embodiment of the present application provides an intelligent answering and identification method for railway production dispatching calls, which obtains the voice information of the railway production dispatching call at the current moment and the information records of each historical railway production dispatching call, and obtains the complexity of the current dispatching call voice information based on the appearance of each keyword in the current dispatching task in the historical dispatching call information records, as well as the confusion of the appearance frequency of all keywords of the current dispatching task. At the same time, the complexity of the dispatching task under the two situations of high and low overall occurrence frequency of events is taken into account, and the rationality of setting the number of keyword retrievals when the complexity of the event complexity is adaptively set is improved; the number of events selected for retrieval is determined based on the complexity of the current dispatching call voice information, so as to determine a more suitable reference historical dispatching record for the current dispatching task for the dispatcher to refer to, thereby avoiding the problem that the final retrieval results cannot be better provided to the dispatcher for reference due to the inability to determine the fine standards for detection, thereby improving the dispatching response efficiency.

[0094] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the above descriptions are of specific embodiments of the present application. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0095] The various embodiments in this application are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0096] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present application shall be included in the scope of protection of the present application.

Claims

1. The method for intelligent answering and identifying incoming calls applied to railway production dispatching is characterized by: The method comprises the following steps: Obtain the voice information of the current railway production dispatch call and the information record of each railway production dispatch call in the past, record them as the current dispatch call voice information and the historical dispatch call information record, and obtain keywords for each; Record the text content of the current scheduled incoming call voice information as the current scheduled task; determine the event occurrence frequency of the current scheduled task based on the occurrence of each keyword in the current scheduled task; Determine the degree of event irrelevance of the current scheduling task based on the occurrence of each keyword in the history of each scheduling call information record; construct the event low-frequency consistency of the current scheduling task based on the disorder of the occurrence frequency of all keywords in the current scheduling task; construct the complexity of the current scheduling task based on the event occurrence frequency, event irrelevance and event low-frequency consistency of the current scheduling task; The preset length of time each day is considered as a cycle; the sensitivity of the peak period of each cycle of the day is calculated based on the number of railway production scheduling calls in each cycle of the day; the number of subsequent event retrievals for the current scheduling task is determined based on the complexity of the current scheduling task, the number of keywords, and the sensitivity during the peak period; Based on the number of subsequent event retrievals, keywords for subsequent retrieval in the current scheduling task are determined.

2. The method for intelligently answering and identifying incoming calls for railway production dispatching according to claim 1, characterized in that: The process of obtaining keywords is as follows: The railway production scheduling knowledge base is used as the input of the keyword extraction algorithm, and the output is the keywords in the scheduling knowledge base; for any word in the scheduling problem of the current scheduling call voice information and the historical scheduling call information records, if the any word is the same as any keyword in the scheduling knowledge base, then the any word is used as the keyword in the scheduling problem corresponding to the current scheduling call voice information or the historical scheduling call information records.

3. The method for intelligently answering and identifying incoming calls for railway production dispatching according to claim 1, characterized in that: The event occurrence frequency is: the average occurrence frequency of all keywords in the current scheduling task.

4. The method for intelligently answering and identifying incoming calls for railway production dispatching according to claim 1, characterized in that: The process of obtaining the degree of irrelevance of the events is as follows: For keyword i in the current scheduling task, obtain the number of times each keyword and keyword i appear together in all historical scheduling call information records, and record it as the first absolute number; the maximum value of all the first absolute numbers of keyword i is taken as the absolute number of occurrences of keyword i; The process of obtaining the relative number of occurrences of each keyword in the current scheduling task is: P1, randomly obtains the first keyword from the current scheduled task and records it as the first keyword; P2, classify the first keyword into the determined keywords; P3, all the remaining keywords except the determined keywords in the current scheduling task will be obtained as undetermined keywords; P4, obtaining the number of co-occurrences of each undetermined keyword and the first keyword in all historical dispatch call information records, recording it as the first relative number; obtaining the maximum value of all the first relative numbers of the first keyword as the relative number of occurrences of the first keyword; recording the undetermined keyword corresponding to the maximum value as the second keyword; P5: Classify the second keyword into the determined keywords and repeat steps P3-P5 until the corresponding relative number of occurrences of each keyword in the current scheduling task is obtained; The degree of irrelevance of the events of the current scheduled task is recorded as W, and the expression of W is: Where I represents the number of keywords in the current scheduling task, w′ i is the relative number of occurrences of the i-th keyword in the current scheduling task, w i The absolute number of occurrences of the i-th keyword in the current scheduled task.

5. The method for intelligently answering and identifying incoming calls for railway production dispatching according to claim 1, characterized in that: The expression of the low-frequency consistency of the event is: E=1-norm(σ q ), where E represents the low-frequency consistency of the event of the current scheduling task, norm() is the normalization function, σ q is the standard deviation of the frequency of occurrence of all keywords in the current scheduling task.

6. The method for intelligently answering and identifying incoming calls for railway production dispatching according to claim 1, characterized in that: The complexity expression of the current scheduling task is: R = Q × W + (1-Q) × E, where R represents the complexity of the current scheduling task, Q represents the frequency of events in the current scheduling task, W represents the degree of irrelevance of the events in the current scheduling task, and E represents the low-frequency consistency of the events in the current scheduling task.

7. The method for intelligently answering and identifying incoming calls for railway production dispatching according to claim 1, characterized in that: The process of obtaining the sensitivity during the peak period of each cycle in a day is: Calculate the ratio of the number of railway production dispatch calls in each cycle of each day to the total number of railway production dispatch calls on that day as the peak proportion of railway production dispatch calls in each cycle of each day; record the sensitivity during the peak period of the kth cycle in a day as U k , U k The expression is: In the formula, norm() is the normalization function, P represents the preset number of days, and x p,k It represents the peak proportion of railway production scheduling calls in the k-th cycle on the p-th day.

8. The method for intelligently answering and identifying incoming calls for railway production dispatching according to claim 1, characterized in that: The expression for the number of subsequent event retrievals for the current scheduled task is: Where J represents the number of subsequent event retrievals for the current scheduling task, I represents the number of keywords in the current scheduling task, R represents the complexity of the current scheduling task, and U′ represents the sensitivity during the peak period of the cycle at the current moment. Indicates the rounding symbol, norm() is the normalization function.

9. The method for intelligently answering and identifying incoming calls for railway production dispatching according to claim 1, characterized in that: The process of determining the keywords for subsequent retrieval in the current scheduling task is as follows: The number of subsequent event retrievals for the current scheduling task is recorded as the first number, all keywords in the current scheduling task are arranged from high to low according to the event occurrence frequency of the keywords to form a keyword sequence, and the first number of keywords in the keyword sequence are used as keywords for subsequent retrieval.

10. An intelligent call answering and identification device for railway production dispatching, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 9 are implemented.

Images