WO2015154641A1 - 一种业务并发性预测方法与预测系统 - Google Patents
一种业务并发性预测方法与预测系统 Download PDFInfo
- Publication number
- WO2015154641A1 WO2015154641A1 PCT/CN2015/075859 CN2015075859W WO2015154641A1 WO 2015154641 A1 WO2015154641 A1 WO 2015154641A1 CN 2015075859 W CN2015075859 W CN 2015075859W WO 2015154641 A1 WO2015154641 A1 WO 2015154641A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- services
- service
- network
- module
- service data
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Definitions
- This paper relates to the field of mobile communication technologies, and in particular to a service concurrency prediction method and prediction system.
- This paper provides a business concurrency prediction method and prediction system, which solves the problem that the related technology is only applicable to the study of the use of a single service.
- a business concurrency prediction method including:
- determining at least two services comprises:
- acquiring the service data corresponding to the at least two services in the at least one historical time period includes:
- the method before the generating the relationship network having the tree structure between the at least two services based on the service data, the method further includes:
- the business data is normalized.
- generating, according to the service data, a relationship network having a tree structure between the at least two services includes:
- generating, according to the service data, a relationship network having a tree structure between the at least two services includes:
- the method further includes:
- a business concurrency prediction system including:
- Determining a module set to: determine at least two services
- An acquiring module configured to: obtain service data corresponding to at least two services determined by the determining module in at least one historical time period;
- a generating module configured to: generate, according to the service data acquired by the acquiring module, a relational network having a tree structure between the at least two services;
- the prediction module is configured to: predict the concurrency of the at least two services according to the relationship network generated by the generating module.
- the determining module is configured to: determine at least two services according to a frequency of use of the service, where the at least two services are at least two common services.
- the acquiring module is configured to: obtain service data corresponding to at least two common services determined by the determining module in at least one historical time period, where the service data is normal. Business data.
- the method further includes:
- the pre-processing module is configured to: normalize the service data acquired by the acquiring module.
- the method further includes:
- the processing module is configured to: obtain, according to the service data acquired by the acquiring module, a mutual relationship between the at least two services;
- the generating module is configured to: generate a relational network having a tree structure between the at least two services according to the mutual relationship obtained by the processing module.
- the processing module includes a calculation sub-module
- the calculating sub-module is configured to calculate a correlation coefficient between the two at least two services according to the service data acquired by the acquiring module;
- the calculating sub-module is further configured to: calculate a distance between the two at least two services according to the correlation coefficient, and use the distance as a mutual relationship between the at least two services;
- the generating module is configured to generate a minimum spanning tree network according to the distance between the at least two services calculated by the calculating submodule, and use the minimum spanning tree network as the at least two services A relational network with a tree structure.
- a computer readable storage medium storing computer executable instructions for performing the method of any of the above.
- the present invention provides a service concurrency prediction method and a prediction system, which generate a relationship network between services based on service data of at least two services in at least one historical time period, and obtain a relationship between the services from the relationship network, thereby Predict the future of the business.
- the business is used as the node, the business distance is constructed as the relationship network, and the hidden relationship between multiple services is visually displayed in a graphical way.
- the original intricate relationship between the services, from which the business concurrency can be effectively predicted.
- the historical data of different regions or different time periods can be flexibly selected to predict the concurrency of the service, which has universal applicability, and the prediction result can provide reference for the service planning and network regulation of the existing network.
- FIG. 1 is a flowchart of a service concurrency prediction method according to Embodiment 1 of the present invention.
- FIG. 3 is a schematic structural diagram of a service concurrency prediction system according to Embodiment 2 of the present invention.
- FIG. 5 is a schematic structural diagram of a minimum spanning tree network according to Embodiment 3 of the present invention.
- FIG. 6 is a partial data set selected from sample data according to Embodiment 4 of the present invention.
- FIG. 7 is a schematic structural diagram of a minimum spanning tree network according to Embodiment 4 of the present invention.
- Embodiment 1 is a diagrammatic representation of Embodiment 1:
- FIG. 1 is a flowchart of a service concurrency prediction method according to Embodiment 1 of the present invention. As shown in FIG. 1 , the service concurrency prediction method includes:
- S101 Determine at least two services
- the service when at least two services are determined, according to the usage of the service, for example, according to the frequency of use of the service, that is, the service whose frequency of use is higher than or equal to the threshold is defined as a common service according to the frequency of use of the service, A service whose frequency is lower than the threshold is defined as a very useful service, and at least two services are determined from all services or part of services, and the determined at least two services are at least two common services, that is, the use of all services or part of services is eliminated.
- the service, the at least two common services are regarded as to be predicted, and the concurrency of the at least two common services is predicted.
- the determined at least two services are determined as at least two services, that is, the determined at least two services are all services or
- the whole service or part of the service is regarded as a service to be predicted, and the concurrency of all or part of the service is predicted.
- S102 Acquire service data corresponding to at least two services in at least one historical time period
- the service data corresponding to the at least two services in the at least one historical time period may be acquired, where the service data includes but is not limited to traffic data of the service, and the source includes but is not limited All base stations in the area to be predicted in the live network.
- the service data when the determined at least two services are at least two common services, acquiring service data corresponding to the at least two common services in at least one historical time period, for example, acquiring the at least two common services.
- the traffic data generated in at least one historical time period, the service data is normal business data, and for abnormal business data, it may be eliminated in the process of acquiring the business data, or may be rejected after the process of acquiring the business data. Eliminated.
- the normal service data and the abnormal service data refer to data at different time points in the service flow sequence.
- the data of individual points deviated significantly from other data, probably due to Acquisition or other unexpected causes of data anomalies, if not eliminated, may have an impact on business correlation calculations.
- actual processing it is considered that if the actual value of a certain point exceeds a specified multiple of the average of the time series, for example, 5 times, it is determined that the point is abnormal business data, and the point value is replaced with the sequence average value during processing. That is, the abnormal business data is business data that exceeds a specified multiple of the time series average.
- the service data when the determined at least two services are all services or part of services, acquiring service data corresponding to the all services or part of services in at least one historical time period, for example, acquiring all services or part of services.
- the traffic data generated in at least one historical time period, the service data is normal business data, and for abnormal business data, it may be eliminated in the process of acquiring the business data, or may be rejected after the process of acquiring the business data. Eliminated.
- the at least one historical time period may be a continuous time period, and the granularity of each historical time period is the same, for example, the granularity is 1 hour, that is, Each historical time period is 1 hour.
- the time span formed by the at least one historical time period may be selected according to actual needs. For example, in order to predict the concurrent situation of a service with other services in a future to-be-predicted time period, the time period to be predicted may be selected. The period before is used as the time span.
- S103 Generate, according to service data, a relational network having a tree structure between at least two services;
- a relational network having a tree structure between at least two services can be generated based on the service data.
- the service data is also normalized, and the embodiment provides a normalization.
- the normalization formula is as shown in the following formula (1.1), and will not be described again for other normalization processing methods.
- x represents the business data of a service in a historical time period. Indicates the average business data of the service over the entire time span, N represents the number of historical time periods, and Z(x) represents the normalized business data.
- a mutual relationship between the at least two services is obtained, and the mutual relationship can reflect the relationship between the services, and according to the mutual relationship,
- a relationship network between the at least two services is generated, the relationship network having a tree structure.
- the tree structure is a hierarchical structure. In the tree structure, no loop is generated between any two nodes, and each path supports bidirectional transmission.
- the structure is characterized by convenient and flexible expansion, low cost and easy promotion, and is suitable for hierarchical management systems of primary or secondary.
- the service data may be arranged according to a time series of at least one historical time period, and according to the service data, a correlation coefficient between at least two services is calculated, Calculation methods include but are not limited to the following methods:
- M is the total duration of the acquired business data, that is, the time span.
- the service data (T 1, 2, 3, ... M) of the service data and the jth service in a historical time period T for the i-th service in a historical time period T.
- a minimum spanning tree network is generated by a Kruskal algorithm according to a distance between at least two services.
- step S205 Deleting the edge, ensuring that the services are not connected to each other when connecting, step S203 is performed;
- the algorithm principle of the Kruskal algorithm is: according to N (N ⁇ 2 and is a positive integer) between the two business Distance, construct a set U, traverse the set U, find the minimum distance value, connect the two services according to the minimum distance value, and then the remaining Continue to find the minimum distance value in the distance, connect the service according to the minimum distance value, and ensure that the services are not connected to each other when connecting, and so on, until all distance values in the set U are traversed, at least two can be obtained.
- N N ⁇ 2 and is a positive integer
- S104 predict the concurrency of at least two services according to the relationship network.
- the concurrency of the service can be predicted according to the relationship network. From the relationship network, such as the minimum spanning tree, the mutual relationship and relationship between the at least two services can be visually seen. The stronger the business, the more likely it is to have concurrency. Therefore, according to the relationship network, it is possible to effectively predict the concurrency of the business in the future.
- a prediction structure is obtained, and the network is optimized according to the prediction result, such as allocating corresponding resources for services in the network.
- the business is used as the node, the business distance is constructed as the relationship network, and the hidden relationship between multiple services is visually displayed in a graphical way.
- the historical data of different regions or different time periods can be flexibly selected to predict the concurrency of the service, which has universal applicability, and the prediction result can provide reference for the service planning and network regulation of the existing network.
- FIG. 3 is a schematic structural diagram of a service concurrency prediction system according to Embodiment 2 of the present invention.
- the service concurrency prediction system includes a determining module 1, an obtaining module 2, a generating module 3, and a prediction module 4, and determining
- the module 1 is configured to determine at least two services
- the obtaining module 2 is configured to acquire service data corresponding to at least two services determined by the determining module 1 in at least one historical time period
- the generating module 3 is configured to be acquired based on the obtaining module 2
- the service data generates a relational network having a tree structure between at least two services
- the prediction module 4 is configured to predict the concurrency of at least two services according to the relationship network generated by the generation module 3.
- the determining module 1 is configured to determine at least two services according to a frequency of use of the service, where the at least two services are at least two common services.
- the obtaining module 2 is configured to acquire service data corresponding to the at least two common services determined by the determining module 2, where the service data is normal service data, in at least one historical time period.
- the preprocessing module 5 is further configured to perform normalization processing on the service data acquired by the obtaining module 2.
- the processing module 6 is further configured to obtain, according to the service data acquired by the obtaining module 2, a mutual relationship between at least two services, and the generating module 3 is configured to obtain a mutual relationship according to the processing module 6. , generating a relational network with a tree structure between at least two services.
- the processing module 6 includes a calculation sub-module 61, and the calculation sub-module 61 is configured to calculate a correlation coefficient between at least two services according to the service data acquired by the acquisition module 2, and further set to calculate at least according to the correlation coefficient.
- the distance between two business pairs the distance is at least two
- the generating module 3 is configured to generate a minimum spanning tree network, and set the minimum spanning tree network as at least two services according to the distance between the two at least two services calculated by the calculating submodule 61.
- a network of relationships between trees are examples of relationships between trees.
- the optimization module 7 further includes: according to the prediction result obtained by the prediction module 4, the optimization module 7 optimizes the network, for example, allocates corresponding resources for services in the LTE (Long Term Evolution) network.
- the optimization module 7 optimizes the network, for example, allocates corresponding resources for services in the LTE (Long Term Evolution) network.
- Embodiment 3 is a diagrammatic representation of Embodiment 3
- the service involved in this embodiment is N (N ⁇ 2 and is a positive integer).
- the service most commonly used by users in the existing network, and the concurrent use of the service most commonly used by users in the predicted area is predicted and acquired in the future.
- At least one historical time period, the service data corresponding to the service most frequently used by the user, the service data is derived from all base stations in the area to be predicted in the live network, and the at least one historical time period is a continuous time period, and each two history
- the interval between time periods is 1 hour, that is, the time granularity is 1 hour, and the time span is 15 consecutive days before the predicted time period.
- FIG. 4 is a partial data set selected from sample data according to Embodiment 3 of the present invention. As shown in FIG. 4, the obtained service data is filtered, and the filtered service data is timed according to at least one historical time period. The sequences are arranged and arranged.
- M is the total duration of the acquired business data, that is, the time span.
- the service data (T 1, 2, 3, ... M) of the service data and the jth service in a historical time period T for the i-th service in a historical time period T.
- a minimum spanning tree network is generated by the network construction method, and the minimum spanning tree network is used as a relational network between at least two services, and the minimum generation
- the tree network has a tree structure.
- the Kruskal algorithm is taken as an example to generate a minimum spanning tree network, according to N (N ⁇ 2 and a positive integer) between two services.
- FIG. 5 is a schematic structural diagram of a minimum spanning tree network according to Embodiment 3 of the present invention.
- FIG. 5 which is a minimum spanning tree network obtained according to service data in a region to be predicted within 15 days, in FIG. 5, multiple The nodes represent different services.
- the node size represents the importance of the node's position in the network. The larger the node, the easier the service is to co-exist with multiple services.
- the color depth of the edge represents the service distance, and the color depth indicates the minimum service distance. That is, the business connected to the dark side is most likely to be concurrent, and the concurrency of any kind of business and other services can be obtained from the figure.
- the concurrency of the N services is predicted, and the relationship between the services is obtained from FIG. 5, and the concurrency of any service with other services in a future period is predicted by the association relationship of the services.
- the minimum spanning tree network of this embodiment it can be predicted that in the future, when the service such as Youku appears, it is likely that there will be four types of services: fast broadcast, LETV, 56VIDEO, and PPLIVE.
- the concurrency of other services can also be visually seen from the network. By obtaining the concurrent situation of the service, the network can be effectively regulated and optimized.
- Embodiment 4 is a diagrammatic representation of Embodiment 4:
- the service involved in this embodiment is N (N ⁇ 2 and is a positive integer) services of all base stations in the to-be-predicted area in the existing network, and the concurrency of all services in the predicted area is predicted in the future period.
- Acquiring service data corresponding to all services in at least one historical time period the service data is derived from all base stations in the area to be predicted in the current network, and the at least one historical time period is a continuous time period, and each of the two historical time periods
- the interval between the two is 1 hour, that is, the time granularity is 1 hour, and the time span is 18 consecutive days before the predicted time period. For example, the time span is from March 8, 2014 to March 26, 2014.
- FIG. 6 is a partial data set selected from sample data according to Embodiment 4 of the present invention. As shown in FIG. 6, the obtained service data is filtered, and the filtered service data is time series according to at least one historical time period. Organize and arrange.
- M is the total duration of the acquired business data, that is, the time span.
- the service data (T 1, 2, 3, ... M) of the service data and the jth service in a historical time period T for the i-th service in a historical time period T.
- a minimum spanning tree network is generated by the network construction method, and the minimum spanning tree network is used as a relational network between at least two services, and the minimum generation
- the tree network has a tree structure.
- the Kruskal algorithm is taken as an example to generate a minimum spanning tree network, according to N (N ⁇ 2 and a positive integer) between two services.
- FIG. 7 is a schematic structural diagram of a minimum spanning tree network according to Embodiment 4 of the present invention.
- a minimum spanning tree network obtained according to service data in a region to be predicted within 18 days, in FIG. 7, multiple
- the nodes represent different services.
- the node size represents the importance of the node's position in the network. The larger the node, the easier it is to concurrently coexist with multiple services. From this figure, the concurrency of any service and other services can be obtained. .
- the concurrent situation of the N services is predicted, and the relationship between the services is obtained from FIG. 7, and the relationship between the services and any other services is predicted by the associated relationship of the services in the future.
- the service such as Youku appears
- the most likely concurrent services are fast broadcast, popular, SOHU-TV, 56VIDEO, PPLIVE, etc.
- the concurrency of several services and other services can also be seen intuitively from the network. By obtaining the concurrency of the services, the network can be effectively regulated and optimized.
- all or part of the steps of the above embodiments may also be implemented by using an integrated circuit. These steps may be separately fabricated into individual integrated circuit modules, or multiple modules or steps may be fabricated into a single integrated circuit module. achieve.
- the devices/function modules/functional units in the above embodiments may be implemented by a general-purpose computing device, which may be centralized on a single computing device or distributed over a network of multiple computing devices.
- the device/function module/functional unit in the above embodiment When the device/function module/functional unit in the above embodiment is implemented in the form of a software function module and sold or used as a stand-alone product, it can be stored in a computer readable storage medium.
- the above mentioned computer readable storage medium may be a read only memory, a magnetic disk or an optical disk or the like.
- the embodiment of the present invention generates a relationship network between the services based on the service data of the at least two services in the at least one historical time period, and obtains the relationship between the services from the relationship network, thereby predicting the future concurrency of the service.
- the business is used as the node, the business distance is constructed as the relationship network, and the hidden relationship between multiple services is visually displayed in a graphical way.
- the original intricate relationship between the services, from which the business concurrency can be effectively predicted.
- the historical data of different regions or different time periods can be flexibly selected to predict the concurrency of the service, which has universal applicability, and the prediction result can provide reference for the service planning and network regulation of the existing network.
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
本文提供了一种业务并发性预测方法及预测系统，该方法包括确定至少两个业务；获取在至少一个历史时间段内，与至少两个业务对应的业务数据；基于业务数据，生成至少两个业务之间具有树型结构的关系网络；根据关系网络，对至少两个业务的并发性进行预测。
Description
本文涉及移动通信技术领域，尤其涉及一种业务并发性预测方法与预测系统。
随着通信技术的飞速发展和终端功能的逐渐丰富，移动网络的数据业务流量迅猛增长，各类移动互联网业务和应用层出不穷。与传统互联网相比，移动互联网下终端用户需求更加多样化和复杂化，这也促使移动互联网由传统的单业务向着多业务平台发展，多业务的出现给运营商的网络运营带来了巨大的冲击。为了提高网络的承载能力，需要科学准确地分析各数据业务之间的并发性。
相关技术对数据业务的分析还大多集中于用户对单一业务使用行为的研究，对于多种数据业务之间的关系研究还未成熟。
发明内容
本文提供了一种业务并发性预测方法及预测系统，解决了相关技术只适用于对单一业务的使用情况进行研究的问题。
一种业务并发性预测方法，包括：
确定至少两个业务；
获取在至少一个历史时间段内，与所述至少两个业务对应的业务数据；
基于所述业务数据，生成所述至少两个业务之间具有树型结构的关系网络；
根据所述关系网络，对所述至少两个业务的并发性进行预测。
在本发明的一种实施例中，确定至少两个业务包括：
根据业务的使用频率，确定至少两个业务，所述至少两个业务为至少两
个常用业务。
在本发明的一种实施例中，获取在至少一个历史时间段内，与所述至少两个业务对应的业务数据包括：
获取在至少一个历史时间段内，与所述至少两个常用业务对应的业务数据，所述业务数据为正常业务数据。
在本发明的一种实施例中，在基于所述业务数据，生成所述至少两个业务之间具有树型结构的关系网络之前，还包括：
对所述业务数据进行归一化处理。
在本发明的一种实施例中，基于所述业务数据，生成所述至少两个业务之间具有树型结构的关系网络包括：
根据所述业务数据，得到所述至少两个业务两两之间的相互关系；
根据所述相互关系，生成所述至少两个业务之间具有树型结构的关系网络。
在本发明的一种实施例中，基于所述业务数据，生成所述至少两个业务之间具有树型结构的关系网络包括：
根据所述业务数据，计算所述至少两个业务两两之间的相关系数；
根据所述相关系数，计算所述至少两个业务两两之间的距离，将所述距离作为所述至少两个业务两两之间的相互关系；
根据所述至少两个业务两两之间的距离，生成最小生成树网络，将所述最小生成树网络作为所述至少两个业务之间具有树型结构的关系网络。
在本发明的一种实施例中，在根据所述关系网络，对所述至少两个业务的并发性进行预测之后，还包括：
根据预测结果，对网络进行优化。
一种业务并发性预测系统，包括：
确定模块，设置为：确定至少两个业务；
获取模块，设置为：获取在至少一个历史时间段内，与所述确定模块确定的至少两个业务对应的业务数据；
生成模块，设置为：基于所述获取模块获取的业务数据，生成所述至少两个业务之间具有树型结构的关系网络；
预测模块，设置为：根据所述生成模块生成的关系网络，对所述至少两个业务的并发性进行预测。
在本发明的一种实施例中，所述确定模块是设置为：根据业务的使用频率，确定至少两个业务，所述至少两个业务为至少两个常用业务。
在本发明的一种实施例中，所述获取模块是设置为：获取在至少一个历史时间段内，与所述确定模块确定的至少两个常用业务对应的业务数据，所述业务数据为正常业务数据。
在本发明的一种实施例中，还包括：
预处理模块，设置为：对所述获取模块获取的业务数据进行归一化处理。
在本发明的一种实施例中，还包括：
处理模块，设置为：根据所述获取模块获取的业务数据，得到所述至少两个业务两两之间的相互关系；
所述生成模块是设置为：根据所述处理模块得到的相互关系，生成所述至少两个业务之间具有树型结构的关系网络。
在本发明的一种实施例中，所述处理模块包括计算子模块；
所述计算子模块设置为：根据所述获取模块获取的业务数据，计算所述至少两个业务两两之间的相关系数；
所述计算子模块还设置为：根据所述相关系数，计算所述至少两个业务两两之间的距离，将所述距离作为所述至少两个业务两两之间的相互关系；
所述生成模块是设置为：根据所述计算子模块计算得到的至少两个业务两两之间的距离，生成最小生成树网络，将所述最小生成树网络作为所述至少两个业务之间具有树型结构的关系网络。
一种计算机可读存储介质，存储有计算机可执行指令，所述计算机可执行指令用于执行上述任一项的方法。
本文提供一种业务并发性预测方法及预测系统，基于至少两个业务在至少一个历史时间段内的业务数据，生成业务之间的关系网络，从该关系网络中得到业务之间的关系，从而预测出业务未来的并发情况。采用社会网络的分析方法，以业务作为节点，业务距离为边构建了关系网络，将多个业务之间隐含的关系以图形化的方式直观地展示出来，化繁为简地呈现出了多个业务之间原本错综复杂的联系，从该业务网络中可以对业务并发情况进行有效预测。此外，可以根据预测需求，灵活地选取不同地区或者不同时间段的历史数据来预测业务的并发情况，具有普遍的适用性，其预测结果可以对现网的业务规划以及网络调控提供参考。
附图概述
图1为本发明实施例一提供的业务并发性预测方法的流程图；
图2为本发明实施例一提供的克鲁斯卡尔(Kruskal)算法的流程图；
图3为本发明实施例二提供的业务并发性预测系统的结构示意图；
图4为本发明实施例三提供的从样本数据中选取的部分数据集合；
图5为本发明实施例三提供的最小生成树网络的结构示意图；
图6为本发明实施例四提供的从样本数据中选取的部分数据集合；
图7为本发明实施例四提供的最小生成树网络的结构示意图。
由于用户社会属性的不同，对于不同业务的使用存在着一定的规律，这就使得我们可以通过对现网数据的分析来预测不同数据业务之间的并发性，从而提前掌握各业务的使用情况，为网络规划、优化、扩容等提供一定指导，从而提高网络对日益丰富的数据业务的承载能力。
下面结合附图对本发明的实施方式做详细说明。
实施例一：
如图1为本发明实施例一提供的业务并发性预测方法的流程图，如图1所示，该业务并发性预测方法包括：
S101：确定至少两个业务；
为了提前掌握多个业务的使用情况并提高网络的承载能力，对一地区未来一段时间内的业务并发性情况进行预测，根据预测需求，合理确定至少两个业务，对于该业务，其包括但不局限于QQ、微信、优酷、土豆、淘宝、微博、百度地图等。
在一些实施例中，当确定至少两个业务时，根据业务的使用情况，例如，根据业务的使用频率，即在一段时间内，将使用频率高于或等于阈值的业务定义为常用业务，使用频率低于阈值的业务定义为非常用业务，从所有业务或部分业务中确定至少两个业务，该确定的至少两个业务为至少两个常用业务，即剔除所有业务或部分业务中的非常用业务，将该至少两个常用业务作为待预测业务，对这至少两个常用业务的并发性进行预测。
在另一些实施例中，当确定至少两个业务时，不需考虑业务的使用情况，而是将所有业务或部分业务确定为至少两个业务，即该确定的至少两个业务为所有业务或部分业务，将该全部业务或部分业务作为待预测业务，对全部业务或部分业务的并发性进行预测。
S102：获取在至少一个历史时间段内，与至少两个业务对应的业务数据；
当确定完成至少两个业务之后，即可获取在至少一个历史时间段内，与该至少两个业务对应的业务数据，该业务数据包括但不局限于业务的流量数据，其来源包括但不局限于现网中待预测地区的所有基站。
在一些实施例中，当确定的至少两个业务为至少两个常用业务时，获取在至少一个历史时间段内，与该至少两个常用业务对应的业务数据，例如获取这至少两个常用业务在至少一个历史时间段内所产生的流量数据，该业务数据为正常业务数据，对于异常业务数据，其可以在业务数据的获取过程中即被剔除，也可以在业务数据的获取过程后再被剔除。
其中，正常业务数据和异常业务数据均指业务流量序列中的不同时间点数据。在前期的处理中发现，个别点的数据明显偏离其他数据，可能是由于
采集或者其他突发原因导致了数据异常，如果不剔除这些数据，可能会对业务的相关性计算产生影响。在实际处理中，认为如果某点实际值超过了时间序列平均值的指定倍数，比如5倍，则确定该点即为异常业务数据，处理中将该点值替换为序列平均值。也即：异常业务数据为超过时间序列平均值的指定倍数的业务数据。在另一些实施例中，当确定的至少两个业务为所有业务或部分业务时，获取在至少一个历史时间段内，与该所有业务或部分业务对应的业务数据，例如获取所有业务或部分业务在至少一个历史时间段内所产生的流量数据，该业务数据为正常业务数据，对于异常业务数据，其可以在业务数据的获取过程中即被剔除，也可以在业务数据的获取过程后再被剔除。
在上述技术方案中，对于上述至少一个历史时间段，可选地，这至少一个历史时间段可为连续时间段，且每个历史时间段的颗粒度均相同，如颗粒度为1小时，即每个历史时间段的时长均为1小时。此外，对于该至少一个历史时间段所组成的时间跨度，可以根据实际需求进行选择，例如，为了预测一业务在未来待预测时间段内与其他业务的并发情况，则可以选择该待预测时间段前的一段时间作为时间跨度。
S103：基于业务数据，生成至少两个业务之间具有树型结构的关系网络；
当获取到业务数据后，即可基于该业务数据，生成至少两个业务之间具有树型结构的关系网络。
在一些实施例中，为了简化计算，在基于业务数据，生成至少两个业务之间具有树型结构的关系网络之前，还对这些业务数据进行归一化处理，本实施例提供一种归一化处理方法，该归一化公式如下式(1.1)所示，对于其他归一化处理方法则不再赘述。
其中，x表示一业务在一个历史时间段内的业务数据，
表示该业务在整个时间跨度内的平均业务数据，N表示历史时间段的数目，Z(x)表示归一化后的业务数据。
在本实施例中，当获取到业务数据后，根据该业务数据，得到这至少两个业务两两之间的相互关系，该相互关系能够反映业务之间的联系，根据该相互关系，即可生成这至少两个业务之间的关系网络，该关系网络具有树型结构。该树型结构是一种分级结构，在树型结构中，任意两个节点之间不产生回路，每条通路都支持双向传输。这种结构的特点是扩充方便、灵活，成本低，易推广，适合于分主次或分等级的层次型管理系统。
在本实施例中，当获取到业务数据后，可以将这些业务数据按照至少一个历史时间段的时间序列进行整理排列，根据该业务数据，计算至少两个业务两两之间的相关系数，其计算方式包括但不局限于以下方式：
对待预测的至少两个业务，根据这至少两个业务的业务数据，计算第i(i＝1,2,3……N)个业务与第j(j＝1,2,3……N)个业务之间的相关系数ρij(i≠j)，其计算公式如下式(1.2)所示：
其中，M为获取的业务数据的总时长，即时间跨度，
分别为第i个业务在时间跨度内的平均业务数据、第j个业务在时间跨度内的平均业务数据，
为第i个业务在一个历史时间段T内的业务数据、第j个业务在一个历史时间段T内的业务数据(T＝1,2,3……M)。
当计算得到相关系数ρij之后，根据该相关系数ρij，计算至少两个业务两两之间的距离，将距离作为至少两个业务两两之间的相互关系，其计算方式包括但不局限于以下方式：
根据相关系数ρij，计算第i个业务与第j个业务之间的距离dij(i≠j)，其计算公式如下式(1.3)所示：
当计算得到至少两个业务两两之间的距离后，根据该距离，通过网络构建方法，生成最小生成树网络，将该最小生成树网络作为至少两个业务之间的关系网络，该最小生成树网络具有树型结构。针对该网络构建方法，任意
最小生成树的算法均适用，其包括但不局限于伪代码法、普里姆(Prim)算法、克鲁斯卡尔(Kruskal)算法、帕斯卡(Pascal)算法，本实施例中以Kruskal算法为例进行说明，其他方法也同样适用。
如图2为本发明实施例一提供的Kruskal算法的流程图，根据至少两个业务两两之间的距离，通过Kruskal算法，生成最小生成树网络。
如图2所示，包括如下步骤：
S201：根据N(N≥2且为正整数)个业务两两之间的
个距离，构建集合U，遍历该集合U；
S202：判断是否全部遍历，如果是，则结束，如果否，则执行S203；
S203：找出其中的最小距离值；
S204：判断业务是否成环，如果是，则执行S205，如果否，则执行S206；
S205：删除此边，保证在连接时，业务之间不连成环，执行步骤S203；
S206：连接此边并从列表中删除，即根据该最小距离值连接这两个业务，执行S201。
从图2可以看出，Kruskal算法的算法原理为：根据N(N≥2且为正整数)个业务两两之间的
个距离，构建集合U，遍历该集合U，找出其中的最小距离值，根据该最小距离值连接这两个业务，然后在剩下的
个距离中继续寻找最小距离值，根据该最小距离值连接业务，同时保证在连接时，业务之间不连成环，如此反复，直至遍历集合U中的所有距离值，即可得到至少两个业务之间具有树型结构的最小生成树网络。
S104：根据关系网络，对至少两个业务的并发性进行预测。
当生成关系网络后，即可以根据该关系网络，对业务的并发性进行预测，从该关系网络中，如最小生成树，可以直观地看出这至少两个业务之间的相互关系，相互关系越强的业务之间越容易出现并发情况，因此，根据该关系网络，可以对未来一段时间内业务的并发性进行有效预测。
在本实施例中，当根据关系网络，对至少两个业务的并发性进行预测后，得到预测结构，根据该预测结果，对网络进行优化，如为网络中的业务分配相应的资源等。
采用社会网络的分析方法，以业务作为节点，业务距离为边构建了关系网络，将多个业务之间隐含的关系以图形化的方式直观地展示出来，化繁为简地呈现出了多个业务之间原本错综复杂的联系，从该业务网络中可以对业务并发情况进行有效预测。此外，可以根据预测需求，灵活地选取不同地区或者不同时间段的历史数据来预测业务的并发情况，具有普遍的适用性，其预测结果可以对现网的业务规划以及网络调控提供参考。
实施例二：
如图3为本发明实施例二提供的业务并发性预测系统的结构示意图，如图3所示，该业务并发性预测系统包括确定模块1、获取模块2、生成模块3以及预测模块4，确定模块1设置为确定至少两个业务，获取模块2设置为获取在至少一个历史时间段内，与确定模块1确定的至少两个业务对应的业务数据，生成模块3设置为基于获取模块2获取的业务数据，生成至少两个业务之间具有树型结构的关系网络，预测模块4设置为根据生成模块3生成的关系网络，对至少两个业务的并发性进行预测。
可选地，确定模块1设置为根据业务的使用频率，确定至少两个业务，该至少两个业务为至少两个常用业务。
可选地，获取模块2设置为获取在至少一个历史时间段内，与确定模块2确定的至少两个常用业务对应的业务数据，该业务数据为正常业务数据。
可选地，还包括预处理模块5，预处理模块5设置为对获取模块2获取的业务数据进行归一化处理。
可选地，还包括处理模块6，处理模块6设置为根据获取模块2获取的业务数据，得到至少两个业务两两之间的相互关系，生成模块3设置为根据处理模块6得到的相互关系，生成至少两个业务之间具有树型结构的关系网络。
可选地，处理模块6包括计算子模块61，计算子模块61设置为根据获取模块2获取的业务数据，计算至少两个业务两两之间的相关系数，还设置为根据相关系数，计算至少两个业务两两之间的距离，将距离作为至少两个
业务两两之间的相互关系，生成模块3是设置为根据计算子模块61计算得到的至少两个业务两两之间的距离，生成最小生成树网络，将最小生成树网络作为至少两个业务之间具有树型结构的关系网络。
可选地，还包括优化模块7，根据预测模块4得到的预测结果，优化模块7对网络进行优化，如为LTE(Long Term Evolution，长期演进)网络中的业务分配相应的资源等。
实施例三：
本实施例中涉及的业务为N(N≥2且为正整数)个现网中用户最常使用的业务，对待预测地区用户最常使用的业务在未来一段时间内的并发性进行预测，获取在至少一个历史时间段，与用户最常使用的业务对应的业务数据，该业务数据来源于现网中待预测地区的所有基站，这至少一个历史时间段为连续时间段，且每两个历史时间段之间的间隔时长为1小时，即时间颗粒度为1小时，时间跨度为待预测时间段前连续15天。
如图4为本发明实施例三提供的从样本数据中选取的部分数据集合，如图4所示，对获取到的业务数据进行筛选，将筛选出的业务数据按照至少一个历史时间段的时间序列进行整理排列。
对整理好的待预测的N个业务，根据这N个业务的业务数据，计算第i(i＝1,2,3……N)个业务与第j(j＝1,2,3……N)个业务之间的相关系数ρij(i≠j)，其计算公式如下式(2.1)所示：
其中，M为获取的业务数据的总时长，即时间跨度，
分别为第i个业务在时间跨度内的平均业务数据、第j个业务在时间跨度内的平均业务数据，
为第i个业务在一个历史时间段T内的业务数据、第j个业务在一个历史时间段T内的业务数据(T＝1,2,3……M)。
根据相关系数ρij，计算第i个业务与第j个业务之间的距离dij(i≠j)，其
计算公式如下式(2.2)所示：
当计算得到至少两个业务两两之间的距离后，根据该距离，通过网络构建方法，生成最小生成树网络，将该最小生成树网络作为至少两个业务之间的关系网络，该最小生成树网络具有树型结构。本实施例以Kruskal算法为例，生成最小生成树网络，根据N(N≥2且为正整数)个业务两两之间的
个距离，构建集合U，遍历该集合U，找出其中的最小距离值，根据该最小距离值连接这两个业务，然后在剩下的
个距离中继续寻找最小距离值，根据该最小距离值连接业务，同时保证在连接时，业务之间不连成环，如此反复，直至遍历集合U中的所有距离值，即可得到至少两个业务之间具有树型结构的最小生成树网络。
如图5为本发明实施例三提供的最小生成树网络的结构示意图，如图5所示，即为根据待预测地区15天内的业务数据得到的最小生成树网络，在图5中，多个节点分别代表不同业务，节点大小表征了该节点在网络中地位的重要性，节点越大，该业务越易与多种业务并发，边的颜色深浅表征了业务距离大小，颜色深表示业务距离最小，即颜色深的边所连接的业务最易出现并发，从该图中可以得到任意一种业务与其他业务的并发情况。
根据该最小生成树网络，预测N个业务的并发情况，从图5中得到业务之间的关联关系，通过业务的关联关系预测出任意一种业务在未来一段时间内与其他业务的并发情况。例如，在本实施例的最小生成树网络中，可以预测出在未来一段时间内，当出现优酷这种业务时，很可能会同时出现快播、LETV(乐视)、56VIDEO、PPLIVE这四类业务，其他业务的并发性也可以从该网络中直观地看出，通过得到业务的并发情况，可以对网络进行有效的调控与优化。
实施例四：
本实施例中涉及的业务为N(N≥2且为正整数)个现网中待预测地区所有基站的业务，对待预测地区所有业务在未来一段时间内的并发性进行预测，
获取在至少一个历史时间段，与所有业务对应的业务数据，该业务数据来源于现网中待预测地区的所有基站，这至少一个历史时间段为连续时间段，且每两个历史时间段之间的间隔时长为1小时，即时间颗粒度为1小时，时间跨度为待预测时间段前连续18天，例如，时间跨度为2014年3月8日至2014年3月26日。
图6为本发明实施例四提供的从样本数据中选取的部分数据集合，如图6所示，对获取到的业务数据进行筛选，将筛选出的业务数据按照至少一个历史时间段的时间序列进行整理排列。
对整理好的待预测的N个业务，根据这N个业务的业务数据，计算第i(i＝1,2,3……N)个业务与第j(j＝1,2,3……N)个业务之间的相关系数ρij(i≠j)，其计算公式如下式(3.1)所示：
其中，M为获取的业务数据的总时长，即时间跨度，
分别为第i个业务在时间跨度内的平均业务数据、第j个业务在时间跨度内的平均业务数据，
为第i个业务在一个历史时间段T内的业务数据、第j个业务在一个历史时间段T内的业务数据(T＝1,2,3……M)。
根据相关系数ρij，计算第i个业务与第j个业务之间的距离dij(i≠j)，其计算公式如下式(3.2)所示：
当计算得到至少两个业务两两之间的距离后，根据该距离，通过网络构建方法，生成最小生成树网络，将该最小生成树网络作为至少两个业务之间的关系网络，该最小生成树网络具有树型结构。本实施例以Kruskal算法为例，生成最小生成树网络，根据N(N≥2且为正整数)个业务两两之间的
个距离，构建集合U，遍历该集合U，找出其中的最小距离值，根据该最小距离值连接这两个业务，然后在剩下的
个距离中继续寻找最小距离值，根据该最
小距离值连接业务，同时保证在连接时，业务之间不连成环，如此反复，直至遍历集合U中的所有距离值，即可得到至少两个业务之间具有树型结构的最小生成树网络。
如图7为本发明实施例四提供的最小生成树网络的结构示意图，如图7所示，即为根据待预测地区18天内的业务数据得到的最小生成树网络，在图7中，多个节点分别代表不同业务，节点大小表征了该节点在网络中地位的重要性，节点越大，该业务越易与多种业务并发，从该图中可以得到任意一种业务与其他业务的并发情况。
根据该最小生成树网络，预测N个业务的并发情况，从图7中得到业务之间的关联关系，通过业务的关联关系预测出任意一种业务在未来一段时间内与其他业务的并发情况。例如，在本实施例的最小生成树网络中，可以预测出在未来一段时间内，当出现优酷这种业务时，最可能会并发的业务是快播、风行、SOHU-TV、56VIDEO、PPLIVE等几种业务，其他业务的并发性也可以从该网络中直观地看出，通过得到业务的并发情况，可以对网络进行有效的调控与优化。
本领域普通技术人员可以理解上述实施例的全部或部分步骤可以使用计算机程序流程来实现，所述计算机程序可以存储于一计算机可读存储介质中，所述计算机程序在相应的硬件平台上(如系统、设备、装置、器件等)执行，在执行时，包括方法实施例的步骤之一或其组合。
可选地，上述实施例的全部或部分步骤也可以使用集成电路来实现，这些步骤可以被分别制作成一个个集成电路模块，或者将它们中的多个模块或步骤制作成单个集成电路模块来实现。
上述实施例中的装置/功能模块/功能单元可以采用通用的计算装置来实现，它们可以集中在单个的计算装置上，也可以分布在多个计算装置所组成的网络上。
上述实施例中的装置/功能模块/功能单元以软件功能模块的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。上述提到的计算机可读取存储介质可以是只读存储器，磁盘或光盘等。
本发明实施例基于至少两个业务在至少一个历史时间段内的业务数据，生成业务之间的关系网络，从该关系网络中得到业务之间的关系，从而预测出业务未来的并发情况。采用社会网络的分析方法，以业务作为节点，业务距离为边构建了关系网络，将多个业务之间隐含的关系以图形化的方式直观地展示出来，化繁为简地呈现出了多个业务之间原本错综复杂的联系，从该业务网络中可以对业务并发情况进行有效预测。此外，可以根据预测需求，灵活地选取不同地区或者不同时间段的历史数据来预测业务的并发情况，具有普遍的适用性，其预测结果可以对现网的业务规划以及网络调控提供参考。
Claims (14)
- 一种业务并发性预测方法，包括：确定至少两个业务；获取在至少一个历史时间段内，与所述至少两个业务对应的业务数据；基于所述业务数据，生成所述至少两个业务之间具有树型结构的关系网络；根据所述关系网络，对所述至少两个业务的并发性进行预测。
- 根据权利要求1所述的业务并发性预测方法，其中，确定至少两个业务包括：根据业务的使用频率，确定至少两个业务，所述至少两个业务为至少两个常用业务，所述常用业务为使用频率高于或等于阈值的业务。
- 根据权利要求2所述的业务并发性预测方法，其中，获取在至少一个历史时间段内，与所述至少两个业务对应的业务数据包括：获取在至少一个历史时间段内，与所述至少两个常用业务对应的业务数据，所述业务数据为正常业务数据。
- 根据权利要求1所述的业务并发性预测方法，在基于所述业务数据，生成所述至少两个业务之间具有树型结构的关系网络之前，还包括：对所述业务数据进行归一化处理。
- 根据权利要求1-4任一项所述的业务并发性预测方法，其中，基于所述业务数据，生成所述至少两个业务之间具有树型结构的关系网络包括：根据所述业务数据，得到所述至少两个业务两两之间的相互关系；根据所述相互关系，生成所述至少两个业务之间具有树型结构的关系网络。
- 根据权利要求5所述的业务并发性预测方法，其中，基于所述业务数据，生成所述至少两个业务之间具有树型结构的关系网络包括：根据所述业务数据，计算所述至少两个业务两两之间的相关系数；根据所述相关系数，计算所述至少两个业务两两之间的距离，将所述距离作为所述至少两个业务两两之间的相互关系；根据所述至少两个业务两两之间的距离，生成最小生成树网络，将所述最小生成树网络作为所述至少两个业务之间具有树型结构的关系网络。
- 根据权利要求1-4任一项所述的业务并发性预测方法，在根据所述关系网络，对所述至少两个业务的并发性进行预测之后，还包括：根据预测结果，对网络进行优化。
- 一种业务并发性预测系统，包括：确定模块，设置为：确定至少两个业务；获取模块，设置为：获取在至少一个历史时间段内，与所述确定模块确定的至少两个业务对应的业务数据；生成模块，设置为：基于所述获取模块获取的业务数据，生成所述至少两个业务之间具有树型结构的关系网络；预测模块，设置为：根据所述生成模块生成的关系网络，对所述至少两个业务的并发性进行预测。
- 根据权利要求8所述的业务并发性预测系统，其中，所述确定模块是设置为：根据业务的使用频率，确定至少两个业务，所述至少两个业务为至少两个常用业务。
- 根据权利要求9所述的业务并发性预测系统，其中，所述获取模块是设置为：获取在至少一个历史时间段内，与所述确定模块确定的至少两个常用业务对应的业务数据，所述业务数据为正常业务数据。
- 根据权利要求8所述的业务并发性预测系统，还包括：预处理模块，设置为：对所述获取模块获取的业务数据进行归一化处理。
- 根据权利要求8-11任一项所述的业务并发性预测系统，还包括：处理模块，设置为：根据所述获取模块获取的业务数据，得到所述至少两个业务两两之间的相互关系；所述生成模块是设置为：根据所述处理模块得到的相互关系，生成所述至少两个业务之间具有树型结构的关系网络。
- 根据权利要求12所述的业务并发性预测系统，其中，所述处理模块包括计算子模块；所述计算子模块设置为：根据所述获取模块获取的业务数据，计算所述至少两个业务两两之间的相关系数；所述计算子模块还设置为：根据所述相关系数，计算所述至少两个业务两两之间的距离，将所述距离作为所述至少两个业务两两之间的相互关系；所述生成模块是设置为：根据所述计算子模块计算得到的至少两个业务两两之间的距离，生成最小生成树网络，将所述最小生成树网络作为所述至少两个业务之间具有树型结构的关系网络。
- 一种计算机可读存储介质，存储有计算机可执行指令，所述计算机可执行指令用于执行权利要求1-7任一项的方法。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410447499.0A CN105426978B (zh) | 2014-09-04 | 2014-09-04 | 一种业务并发性预测方法与预测系统 |
| CN201410447499.0 | 2014-09-04 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2015154641A1 true WO2015154641A1 (zh) | 2015-10-15 |
Family
ID=54287321
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2015/075859 WO2015154641A1 (zh) | 2014-09-04 | 2015-04-03 | 一种业务并发性预测方法与预测系统 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN105426978B (zh) |
| WO (1) | WO2015154641A1 (zh) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114132204A (zh) * | 2021-11-26 | 2022-03-04 | 国网北京市电力公司 | 充电桩维护方法、装置、非易失性存储介质及处理器 |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112686483B (zh) * | 2019-10-17 | 2024-08-13 | 中国移动通信集团陕西有限公司 | 预警区域识别方法、装置、计算设备及计算机存储介质 |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101321178A (zh) * | 2008-07-18 | 2008-12-10 | 中兴通讯股份有限公司 | 一种门户工作流引擎系统及其实现方法 |
| CN102073588A (zh) * | 2010-12-28 | 2011-05-25 | 北京邮电大学 | 一种基于代码静态分析的多线程死锁检测方法及系统 |
| CN103729289A (zh) * | 2013-11-29 | 2014-04-16 | 北京广利核系统工程有限公司 | 一种利用图式标明hpd逻辑可靠性的方法 |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR100347456B1 (ko) * | 1999-08-10 | 2002-08-03 | 학교법인 한국정보통신학원 | 분산 가상 환경에서의 확장성을 위한 예측 기반 동시성 제어방법 |
| CN100501695C (zh) * | 2006-12-25 | 2009-06-17 | 中国电信股份有限公司 | 应用软件在生产环境下的性能预测方法 |
| US20100281469A1 (en) * | 2009-04-30 | 2010-11-04 | Nec Laboratories America, Inc. | Symbolic predictive analysis for concurrent programs |
| CN102457878B (zh) * | 2010-10-18 | 2014-06-04 | 中国移动通信集团四川有限公司 | 一种在关键业务预测中剔除异常数据的方法和装置 |
| CN103136113B (zh) * | 2011-11-25 | 2016-01-13 | 中国科学院沈阳计算技术研究所有限公司 | 面向多核处理器的共享Cache冲突预测方法 |
| CN103186834A (zh) * | 2011-12-31 | 2013-07-03 | 阿里巴巴集团控股有限公司 | 业务流程配置方法和装置 |
| CN103841041B (zh) * | 2012-11-23 | 2018-12-28 | 中兴通讯股份有限公司 | 一种多流业务并发传输控制方法及装置 |
| CN103490956A (zh) * | 2013-09-22 | 2014-01-01 | 杭州华为数字技术有限公司 | 基于业务量预测的自适应节能控制方法及设备、系统 |
| CN103678514B (zh) * | 2013-11-26 | 2017-01-18 | 科大讯飞股份有限公司 | 一种业务趋势预测方法及系统 |
-
2014
- 2014-09-04 CN CN201410447499.0A patent/CN105426978B/zh active Active
-
2015
- 2015-04-03 WO PCT/CN2015/075859 patent/WO2015154641A1/zh active Application Filing
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101321178A (zh) * | 2008-07-18 | 2008-12-10 | 中兴通讯股份有限公司 | 一种门户工作流引擎系统及其实现方法 |
| CN102073588A (zh) * | 2010-12-28 | 2011-05-25 | 北京邮电大学 | 一种基于代码静态分析的多线程死锁检测方法及系统 |
| CN103729289A (zh) * | 2013-11-29 | 2014-04-16 | 北京广利核系统工程有限公司 | 一种利用图式标明hpd逻辑可靠性的方法 |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114132204A (zh) * | 2021-11-26 | 2022-03-04 | 国网北京市电力公司 | 充电桩维护方法、装置、非易失性存储介质及处理器 |
| CN114132204B (zh) * | 2021-11-26 | 2023-11-03 | 国网北京市电力公司 | 充电桩维护方法、装置、非易失性存储介质及处理器 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105426978B (zh) | 2021-05-04 |
| CN105426978A (zh) | 2016-03-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11269834B2 (en) | Detecting quasi-identifiers in datasets | |
| US10915508B2 (en) | Data linking | |
| US8626835B1 (en) | Social identity clustering | |
| RU2607621C2 (ru) | Способ, система и машиночитаемый носитель данных для группирования в социальных сетях | |
| WO2017045472A1 (zh) | 资源预测方法、系统和容量管理装 | |
| WO2015085948A1 (en) | Method, device, and server for friend recommendation | |
| US9774654B2 (en) | Service call graphs for website performance | |
| EP2975538B1 (en) | Computer-implemented method and apparatus for determining relevance of a node in a network | |
| Nasridinov et al. | A decision tree-based classification model for crime prediction | |
| Lujic et al. | Efficient edge storage management based on near real-time forecasts | |
| CN113254472B (zh) | 一种参数配置方法、装置、设备及可读存储介质 | |
| US10482268B1 (en) | Systems and methods for access management | |
| CN114116065A (zh) | 获取拓扑图数据对象的方法、装置、及电子设备 | |
| US10250550B2 (en) | Social message monitoring method and apparatus | |
| CN112085087A (zh) | 业务规则生成的方法、装置、计算机设备及存储介质 | |
| CN113099475A (zh) | 网络质量检测方法、装置、电子设备及可读存储介质 | |
| US20160125095A1 (en) | Lightweight temporal graph management engine | |
| CN111177481A (zh) | 用户标识映射方法及装置 | |
| WO2015154641A1 (zh) | 一种业务并发性预测方法与预测系统 | |
| CN108696418B (zh) | 一种社交网络中隐私保护方法及装置 | |
| WO2019028149A1 (en) | SYSTEMS AND METHODS FOR STRUCTURE EXTRACTION FROM DENSED, LOUD, AND LARGE DIMENSIONS | |
| CN113779017A (zh) | 数据资产管理的方法和装置 | |
| CN108111968B (zh) | 一种基于泛化的位置隐私保护方法 | |
| KR101256922B1 (ko) | 예측 및 분석을 위한 분산 의사 결정 트리 생성방법 | |
| US20220342887A1 (en) | Predictive query processing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 15777171 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 15777171 Country of ref document: EP Kind code of ref document: A1 |