Method and system for searching specific records in big data

Abstract

The invention discloses a search method and system for specific records in big data, and belongs to the technical field of big data. According to the method, feature information such as terminal numbers, calling time and service types is collected, and multi-level filtering strategies such as terminal number preliminary screening, time threshold filtering, calling and called type distinguishing, calling service type and abnormal phenomenon classification screening, geographic position matching and terminal behavior pattern analysis are adopted. And specific records are accurately positioned from massive terminal call records of the mobile communication network. According to the scheme disclosed by the invention, a wireless communication principle and a terminal behavior rule are fused, auxiliary calculation is carried out by utilizing the terminal public data, and a certain piece of call record data under a specific condition is gradually searched by formulating a specific layer-by-layer filtering skill and method in massive terminal call record data; the problems of long time consumption and high cost of manual screening in the prior art are solved, and the method is suitable for a 2G-6G all-standard mobile communication network.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of big data, and particularly relates to a search method and system for specific records in big data. BACKGROUND

[0002] Big data, also known as massive data or huge data, refers to data of a scale so large that it cannot be captured, managed, processed, and arranged into data to assist decision-making within a reasonable time through mainstream software tools. Big data has the characteristics of large volume, high speed, variety, low value density, and veracity.

[0003] User call records are complete connection process data recorded by a cellular mobile communication network when a terminal initiates a service. User call records have various data types, such as call detailed records (CDRs), measurement reports (MRs), and communication history reports (CHRs) in an access network. Call detailed records (CDRs) are used to record key historical information of a call, including terminal characteristic information, call establishment characteristic information, QoS-related information, call process behavior information, and call release-related information. A call detailed record is generated when a user establishes a connection and switches to a new 5G / 6G communication system. Measurement reports (MRs) are mainly generated by a terminal in a wireless resource management process. User call record data is a typical big data. Communication history reports (CHRs) are log files of a base station reflecting problems. CHR logs are log files of a base station used to record problems occurring during a call.

[0004] A search method is a method of using the high performance of a computer to exhaustively search for some or all possible cases of a problem solution space, so as to find the solution to the problem. The search scale is usually reduced according to the conditions before searching, pruning is performed according to the constraints of the problem, and intermediate solutions during the search process are used to avoid repeated calculations, such as enumeration algorithms, depth-first search, Monte Carlo tree search, and other algorithms.

[0005] With the development of mobile Internet, mobile communication networks have generated massive user call record big data. User call record data is used to record key historical information of a call in a mobile network, including terminal characteristic information, call establishment characteristic information, QoS-related information, call process behavior information, and call release-related information. A call detailed record is generated when a user establishes a wireless resource and switches to a new 5G / 6G communication system. In the big data, searching for a specific record, the call record data of the terminal becomes a demand often encountered in the big data application. For example, in the customer service support scene of the telecommunications, when an engineer is processing a complaint of a user, the engineer first needs to search for the call record data of a specific call of the complaint user in the massive user call record data of the whole network; and then analyzes the complaint reason according to the specific user record data. At present, the main search method is to manually search for specific call record data in massive data according to business experience. This method is time-consuming and inefficient, and it is difficult to meet the real-time or quasi-real-time telecommunications application, and cannot be popularized to the positioning of the real-time optimization application scene of the ordinary user. SUMMARY

[0006] The technical purpose of the present application is to solve the problem of long time consumption and low efficiency in the method of searching for a specific record in massive user call record data, which is difficult to meet the real-time or quasi-real-time complaint preprocessing scene, has limited use scene, and cannot be popularized to the positioning of the real-time optimization application scene of the ordinary user. Therefore, a search method and system for a specific record in big data are provided.

[0007] To achieve the above technical purpose, the present application adopts the following technical scheme.

[0008] In a first aspect, the present application provides a search method for a specific record in big data, comprising: collecting feature information related to a target call record, the feature information at least including a terminal number, a call occurrence time, a main / called type of the call, a call service type, an abnormal phenomenon classification, and a call occurrence location; performing data layer-by-layer filtering based on the terminal number, the call occurrence time, the call service type, the abnormal phenomenon classification, and the call occurrence location to obtain the target call record.

[0009] In a second aspect, the present application provides a search system for a specific record in big data, comprising: a feature collection module configured to collect feature information related to a target call record, the feature information at least including a terminal number, a call occurrence time, a main / called type of the call, a call service type, an abnormal phenomenon classification, and a call occurrence location; a multi-layer filtering module configured to perform data layer-by-layer filtering based on the terminal number, the call occurrence time, the call service type, the abnormal phenomenon classification, and the call occurrence location to obtain the target call record.

[0010] In a third aspect, the present application provides a computer readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the search method for a specific record in big data provided by any possible implementation manner of the first aspect.

[0011] In a fourth aspect, the embodiments of the present application provide an electronic device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor implements the method for searching a specific record in big data according to any possible implementation manner of the first aspect when executing the program.

[0012] Compared with the prior art, the method and system for searching a specific record in big data provided by the embodiments of the present application have the following beneficial technical effects: The present scheme changes the traditional way of manually searching for a specific call record in massive data according to business experience, forms a systematic screening strategy based on the creative logical combination of communication, geographic information, and user behavior, and realizes automatic screening through a multi-level filtering strategy, thereby greatly shortening the search time and meeting the real-time or quasi-real-time telecommunication application requirements, such as quickly positioning the relevant call record when an engineer handles a user complaint in a customer service support scenario. Without the need to increase special auxiliary equipment and manual input, the method of fusing business rules and big data distribution rules can maintain rapid, efficient, accurate, and low-cost development and deployment of applications. The method is suitable for mobile terminal location positioning of 2G / 3G / 4G / 5G / 5G-A / 6G and other mobile communication networks, and has the versatility of multiple mobile communication systems. BRIEF DESCRIPTION OF DRAWINGS

[0013] The drawings described herein are for illustrative purposes only and are not intended to limit the scope of the present application in any way. In addition, the shapes, proportions, and sizes of the components in the drawings are only illustrative and are used to help understand the present application, and are not specific limitations on the shapes, proportions, and sizes of the components. Those skilled in the art can select various possible shapes, proportions, and sizes to implement the present application according to specific circumstances under the guidance of the present application. In the drawings: Figure 1 A flowchart of the method for searching a specific record in big data provided by the embodiments of the present application; Figure 2 A single-step detailed flowchart of the method for searching a specific record in big data provided by the embodiments of the present application; Figure 3 A structural schematic diagram of the system for searching a specific record in big data provided by the embodiments of the present application. DETAILED DESCRIPTION

[0014] In order for those skilled in the art to better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative efforts should fall within the scope of protection of the present application.

[0015] Existing search technologies often retrieve multiple records from massive amounts of user call logs for the user to choose from; or they rely on manual methods to find specific call logs based on business experience. These methods are time-consuming and costly, making it difficult to meet the needs of real-time or near-real-time complaint pre-processing scenarios. Their application scenarios are limited, and they cannot be widely adopted for real-time optimization applications for ordinary users.

[0016] This invention integrates business patterns and big data distribution patterns. For example, based on the regularity of wireless communication principles and terminal behavior in mobile communication networks, it uses publicly available terminal data (such as terminal numbers) for auxiliary calculations. By formulating specific layer-by-layer filtering techniques and methods (such as business type, GIS location matching, terminal behavior, etc.) in massive terminal call record data, it gradually searches for a specific call record data under specific conditions.

[0017] The following description, in conjunction with the accompanying drawings and specific embodiments, provides further details.

[0018] Example 1: A method for searching specific records in big data, such as... Figure 1 As shown, it includes the following steps: Step a1: Collect feature information related to the target call record. The feature information includes at least the terminal number, call time, call service type, anomaly classification, and call location. Step a2: Based on the terminal number, call time, call service type, anomaly classification, and call location, perform layer-by-layer data filtering to obtain the target call record; Step a3: When only one record remains in the final filtering result, it is identified as the target specific record.

[0019] In some embodiments, step a2 includes the following steps: Step a2.1: Based on the terminal number, perform preliminary screening on massive terminal call record data to obtain a candidate record set associated with the terminal number; Step a2.2: Filter the candidate record set according to the call occurrence time by setting a time threshold, and retain the records within the time threshold range; Step a2.3: Further filter the filtered records based on the caller and called party types to distinguish between caller and called party records; Step a2.4: Categorize and filter records according to call service type; Step a2.5: Further filter specific terminal call record data according to whether there are any abnormalities in the terminal calls; Step a2.6, abnormal state filtering of the records according to the abnormal phenomenon classification, retaining the records meeting the abnormal phenomenon classification or removing the abnormal records; Step a2.7, geographic positioning of the filtered records, and matching calculation with the geographic positioning result of the call occurrence place, further narrowing the range according to a preset error threshold; Step a2.8, final filtering of the records based on the terminal behavior mode, the terminal behavior mode including: when there are multiple continuous call records in a similar time period, if they are all abnormal records of the same type, the first one is taken; if the abnormal types are different, the first one is screened according to the abnormal phenomenon classification in the feature information; if the abnormal and normal are alternated, the first abnormal record is taken; In the embodiment, the feature information related to the target call record is collected, and the feature information is used as necessary information to implement the search algorithm for specific call record data in big data.

[0020] In some embodiments, the feature information is, for example, terminal public data provided by a user to a telecom operator in a mobile communication network.

[0021] Among them: (1) Call service type: such as data, 5G message, voice, short message (2) Abnormal phenomenon classification: such as caller call failure, called call failure, caller and called call abnormality, disconnection, call drop, paging failure, short message failure, etc.

[0022] The method provided by the application adopts a search algorithm for specific call record data in big data, and performs layer-by-layer filtering on a large amount of terminal call record data according to necessary data to obtain specific call record data.

[0023] In the embodiment, step a2.1 preliminarily screens a group of specific call records from a large amount of user call record data in a mobile cellular network according to the "terminal number". In the large amount of terminal call data of the telecom, the call record data of the specific terminal number is screened out.

[0024] Step a2.2 further filters the specific terminal call record according to the call occurrence time on the basis of step a2.1. On the basis of step a2.1, a time threshold is further set, and a group of terminal call record data corresponding to the time period within the threshold range before and after the problem occurrence time is screened out.

[0025] Step a2.3 further filters the specific terminal call record data according to whether the call initiated by the terminal is a caller or a called party on the basis of the previous step a2.1. On the basis of step a2.1, the caller call record data or the called call record data of the terminal is further filtered according to whether the terminal initiates a call as a caller or a called party.

[0026] Step a2.4 builds upon the previous step by further filtering specific terminal call record data according to the service type of the call initiated by the terminal.

[0027] Based on step a2.3, the call record data of the terminal is further filtered according to whether the terminal initiates a data service, 5G messaging service, voice service or short message service.

[0028] Step a2.5 builds upon the previous step by further filtering specific terminal call record data based on whether there are any abnormalities in the terminal calls.

[0029] If the terminal call is successfully connected and released, then based on step a2.4, abnormal calls are removed, normal calls are retained, and the terminal call record data is further filtered; if there are abnormal phenomena in the terminal call, then based on step a2.4, normal calls are removed, abnormal calls are retained, and the terminal call record data is further filtered.

[0030] Step a2.6, based on the previous step, filters specific terminal call record data according to the abnormal phenomena of the abnormal calls. Building upon step a2.5, further filters terminal call record data based on the abnormal phenomenon classification (e.g., caller initiation failure, called party initiation failure, abnormal call initiation by both parties, dropped call, dropped call, paging failure, SMS failure, etc.).

[0031] Step a2.7, based on the previous step, performs matching calculations on the filtered terminal call record data according to geographical location, further narrowing down the scope of specific terminal call record data. Based on step a2.6, the filtered terminal call record data is geolocated, and the "call location" in the necessary terminal information is also geolocated. Matching calculations are then performed based on the location areas of both to further filter the terminal call record data.

[0032] Step a2.8 builds upon the previous step by further filtering specific terminal call record data based on terminal behavior patterns.

[0033] Terminal behavior patterns indicate that when users encounter abnormal calls, they will usually call again or multiple times from the same location, generating two or more consecutive call records within a similar time period.

[0034] In some embodiments, such as Figure 2 As shown, based on step a2.7, the terminal call records are further filtered into the following three cases: ① If two or more consecutive terminal call records are both abnormal terminal call records and the abnormal call types are the same, then the first call record will be filtered out. ② If two or more consecutive terminal call record data are all abnormal terminal call record data but their abnormal call type (for example, after the terminal disconnects due to abnormal data service, the reconnection may become a call establishment failure), then the first record in the corresponding call record data will be selected according to the specific problem classification provided by the terminal in the necessary information. ③ If two or more consecutive terminal call records show alternating abnormal and normal phenomena (for example, a terminal call is dropped when the elevator door closes, and the call is re-established normally after the door opens), then the first call record showing the abnormal phenomenon is taken as the specific call record data.

[0035] If only one terminal call record is found after the search algorithm, it is considered a specific record. This specific record can then be automatically entered into subsequent AI intelligent analysis.

[0036] The following embodiments, using urban 2G wireless networks as an example, provide a method for searching specific records in big data, including: 1. The first step in searching specific call record data is to collect feature information: Complaint information source: various fields of user complaint data, including: terminal number, time of problem occurrence, calling / called, service type, problem phenomenon category, and location of problem occurrence.

[0037] in: 1) Service types: Voice, 1x data, SMS; 2) Problem categories: calling failure, called failure, abnormal calling / called process, dropped call, paging failure, SMS failure.

[0038] 2. The second step in specific call record data mining is the method of automatically searching for specific call record data in big data using algorithms: 1) Perform preliminary filtering of terminal call record data based on "terminal number": Filter out terminal call record data where the Subscriber number field = "terminal number"; 2) Further filter terminal call log data based on the time the problem occurred: Set a time threshold to filter a set of terminal call record data within the threshold range before and after the time the problem occurred. The time threshold is adjustable (e.g., 15 minutes or 60 minutes). 3) Further filter terminal call record data by caller / called party.

[0039] Originated Digits<>0 means the terminal number is the caller, Calling Party ID<>0 means the terminal number is the callee, also can be represented by Received from Other MSC=2 means the terminal number is the caller, Received from Other MSC=3 means the terminal number is the callee.

[0040] PS: All the terminal call record data of the terminal number of the failed paging = the caller number, no terminal number of the callee number.

[0041] 4) Further filter the terminal call record data according to Allocation Service Option; Alloc.Service=1,3,17,56,68,32768 voice service; Alloc.Service=33 1x data service; Alloc.Service=6,14 short message.

[0042] 5) Remove normal calls and keep abnormal calls to further filter the terminal call record data; Remove normal call terminal call record data of CFC=1 and CFCQ<>8.

[0043] 6) Filter the terminal call record data in the corresponding abnormal call terminal call record data group according to the problem phenomenon classification, service type and caller-callee situation.

[0044] Voice and 1x data abnormal call terminal call record data group: Caller call failure: CFC=3,5,7,9,11,15,17; Callee call failure: CFC=4,6,8,10,12,16,18; Unknown caller-callee call failure: CFC=13; Call drop: CFC=2; Paging failure: CFC=1 and CFCQ=8; Short message abnormal call terminal call record data group: Short message sending failure: CFC=14; 7) GIS positioning of the filtered terminal call record data or terminal call record data group, and GIS positioning of the "problem occurrence location" in the complaint information, and further filtering of the terminal call record data according to the matching of the positioning areas of the two.

[0045] a) "Problem occurrence location" positioning method: GIS positioning according to road name and house number; b) Location method based on terminal call record data: triangulation algorithm; c) Rule: When the "location of the problem" is a clear street name and house number, the terminal call record data is considered to match if the error between the positioning algorithm and the latitude and longitude of the house number is within the following range: inner ring <= 400 meters, inner ring to outer ring <= 600 meters, outside the outer ring <= 800 meters (the error range threshold can be set).

[0046] 8) Further filter terminal call record data based on terminal behavior patterns: Terminal behavior patterns indicate that when a terminal encounters an abnormal call, it may dial again or multiple times from the same location, generating two or more consecutive terminal call record data within a similar time period (threshold can be set, default is within 3 minutes). At this time, the terminal call record data is further filtered into the following three situations: a) If two or more consecutive terminal call record data are both abnormal terminal call record data and their CFC and CFCQ are consistent, then the first one is filtered out; b) If two or more consecutive terminal call record data are abnormal terminal call record data but their CFC and CFCQ are inconsistent (e.g., redialing after a dropped call may become a call setup failure), then filter according to the specific problem category of the complaint, or list them all (the default is the filtering method). c) If two or more consecutive terminal call records show alternating abnormal and normal phenomena (such as a dropped call caused by the elevator door closing, and the call being re-established normally after the door opens), then the abnormal terminal call record data shall be taken.

[0047] 3. If only one terminal call record is found after the search algorithm, it is considered a specific record. This specific record can automatically proceed to subsequent AI intelligent analysis.

[0048] This invention proposes a method for searching specific call records in large amounts of terminal call data (such as massive amounts of data in mobile communication networks). It has universality across multiple mobile communication standards and is applicable to the location positioning of mobile terminals in various mobile communication networks such as 2G / 3G / 4G / 5G / 5G-A / 6G, etc.

[0049] To facilitate understanding of this invention, a 2G network is used as an example for search calculations. However, it should be noted that this invention pertains to the field of mobile communications, specifically a method for searching specific data based on mobile communication principles and auxiliary data. It is not limited to 2G networks; this positioning method is also applicable to searching call record data of various mobile terminals, including 3G / 4G / 5G / 5G-A / 6G, etc. Therefore, the scope of protection of this invention should include the mobile communication field across 2G / 3G / 4G / 5G / 5G-A / 6G, etc., and not just the 2G network exemplified in the embodiments.

[0050] Based on the same inventive concept as the search method for specific records in big data provided in the above embodiments, this application also provides a search system for specific records in big data, such as... Figure 3 As shown, it includes a feature acquisition module and a multi-layer filtering module. The feature acquisition module is used to collect feature information related to the target call record. The feature information includes at least the terminal number, call occurrence time, caller / called type, call service type, anomaly classification, and call location. The multi-layer filtering module is used to filter the data layer by layer based on the terminal number, call occurrence time, call service type, anomaly classification, and call location to obtain the target call record.

[0051] In this embodiment, the multi-layer filtering module is specifically used for: preliminary screening of massive terminal call record data based on the terminal number to obtain a candidate record set associated with the terminal number; filtering the candidate record set according to a time threshold set based on the call occurrence time, retaining records within the time threshold range; further filtering the filtered records based on the caller / callee type to distinguish between caller and callee records; classifying and filtering records according to call service type; further filtering specific terminal call record data based on whether there are abnormal phenomena in the terminal call; filtering records for abnormal states based on the abnormal phenomenon classification, retaining records that meet the abnormal phenomenon classification or removing abnormal records; geolocating the filtered records and matching them with the geolocation results of the call occurrence location, further narrowing the range according to a preset error threshold; and finally filtering the records based on the terminal behavior pattern, which includes: when there are multiple consecutive call records within a similar time period, if they are all abnormal records of the same type, the first one is selected; if the abnormal types are different, the first one is selected according to the abnormal phenomenon classification in the feature information; if abnormal and normal alternate, the first abnormal record is selected; when the final filtering result has only one record, it is determined as the target call record.

[0052] The search system for specific records in big data provided in this application uses a multi-layered, progressive filtering approach, considering factors such as terminal number, time, caller / called type, service type, anomaly classification, geographical location, and terminal behavior patterns, to gradually narrow down the search from massive amounts of data and ultimately pinpoint a specific call record. It integrates knowledge of the principles of wireless communication in mobile networks and the regularities of terminal behavior, utilizes publicly available terminal data for auxiliary calculations, and develops layered filtering techniques and methods to improve the accuracy and targeting of the search.

[0053] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the search method for specific records in big data as described above.

[0054] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the search method for specific records in big data as described above.

[0055] This invention utilizes the principles of wireless communication and the regularities of terminal behavior to gradually search for a specific call record based on specific conditions by employing layered filtering techniques and methods from massive amounts of terminal call record data. At the same time, the method of this invention maintains the speed, accuracy, and low cost of developing and deploying applications, meeting the corresponding real-time or near-real-time applications in the telecommunications industry (such as complaint preprocessing), without adding any specialized auxiliary equipment or manpower.

[0056] The above provides a detailed description of a method and system for searching specific records in big data provided by this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the concept of this application and should not be construed as a limitation on the scope of protection of this application.

Claims

1. A method for searching specific records in big data, characterized in that, include: Collect feature information related to the target call record. The feature information includes at least the terminal number, call time, caller / called type, call service type, abnormal phenomenon classification, and call location. Based on the terminal number, call time, call service type, anomaly classification, and call location, data is filtered layer by layer to obtain the target call record.

2. The method for searching specific records in big data according to claim 1, characterized in that, Based on the terminal number, call time, call service type, anomaly classification, and call location, data is filtered layer by layer to obtain the target call records, including: Based on the terminal number, a preliminary screening is performed on a massive amount of terminal call record data to obtain a set of candidate records associated with the terminal number; The candidate record set is filtered according to a time threshold set based on the call occurrence time, and records within the time threshold range are retained; The filtered records are further filtered based on the caller and callee types to distinguish between caller and callee records. The records are categorized and filtered according to the call service type. Further filter specific terminal call record data based on whether there are any abnormalities in the terminal call; The records are filtered for abnormal states based on the abnormal phenomenon classification, retaining records that match the abnormal phenomenon classification or removing abnormal records. The filtered records are geolocated and matched with the geolocation results of the call location. The range is further narrowed down based on a preset error threshold. The records are finally filtered based on the terminal behavior pattern, which includes: when there are multiple consecutive call records within a similar time period, if they are all abnormal records of the same type, the first one is selected; if the abnormal types are different, the first one is selected according to the abnormal phenomena in the feature information; if abnormal and normal alternate, the first abnormal record is selected. When only one record remains in the final filtering result, it is identified as the target call record.

3. The method for searching specific records in big data according to claim 2, characterized in that, In the matching calculation, the triangulation algorithm is used to locate the terminal call records in the candidate record set, and the location of the call is located using the street name and house number in the geographic information system.

4. The method for searching specific records in big data according to claim 2, characterized in that, The time threshold for similar time periods is set to 3 minutes by default, and can be customized.

5. The method for searching specific records in big data according to claim 2, characterized in that, The time threshold can be dynamically adjusted according to the business scenario, and the default setting is 15 minutes or 60 minutes before or after the time the problem occurs.

6. A search system for specific records in big data, characterized in that, include: The feature acquisition module is used to collect feature information related to the target call record. The feature information includes at least the terminal number, call time, caller / called type, call service type, abnormal phenomenon classification, and call location. The multi-layer filtering module is used to filter data layer by layer based on terminal number, call time, call service type, abnormal phenomenon classification, and call location to obtain target call records.

7. The search system for specific records in big data according to claim 6, characterized in that, The multi-layer filtering module is specifically used for: performing preliminary screening on massive terminal call record data based on the terminal number to obtain a candidate record set associated with the terminal number; filtering the candidate record set according to a time threshold set based on the call occurrence time, and retaining records within the time threshold range; The filtered records are further filtered based on the caller and callee types to distinguish between caller and callee records; the records are also categorized and filtered according to the call service type; and specific terminal call record data is further filtered based on whether there are any abnormalities in the terminal call. The records are filtered for abnormal states based on the abnormal phenomenon classification, retaining records that match the abnormal phenomenon classification or removing abnormal records; the filtered records are geolocated and matched with the geolocation results of the call location, and the range is further narrowed according to a preset error threshold. The records are finally filtered based on the terminal behavior pattern, which includes: when there are multiple consecutive call records within a similar time period, if they are all abnormal records of the same type, the first one is selected. If the abnormality types are different, the first record is selected according to the abnormal phenomenon classification in the feature information; if abnormal and normal alternate, the first abnormal record is selected; when only one record remains in the final filtering result, it is determined to be the target call record.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the search method for specific records in big data as described in any one of claims 1-5.

9. 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, it implements the search method for specific records in big data as described in any one of claims 1-5.

Images