Event prediction method and information processing device applied to event prediction

Abstract

An event prediction method according to an embodiment of the present disclosure includes: receiving one or more observation data related to a prediction event, and analyzing the one or more observation data based on an event prediction model and outputting an event level at which the prediction event occurs, wherein the event prediction model includes a hierarchical analysis network, the hierarchical analysis network includes multiple types of prior events, each type of prior event includes one or more characterizing events, and each type of prior event corresponds to an event level at which the prediction event occurs. An information processing device for event prediction according to an embodiment of the present disclosure further includes a processor configured to execute the above event prediction method.

Description

Technical Field

[0001] Example embodiments of the present disclosure generally relate to the field of computer information, and particularly to an event prediction method and an information processing device for event prediction. Background Art

[0002] Event prediction is applied to various fields in the industrial and industrial sectors, such as power, network fault prediction, energy distribution prediction, production decision-making, transportation, environmental monitoring, and military, etc.

[0003] However, existing event prediction methods simply rely on the collected monitoring data and historical information, and do not comprehensively consider the relationships between the data and the historical information, resulting in insufficient current event prediction capabilities. Summary of the Invention

[0004] In a first aspect of the present disclosure, there is provided a method for event prediction, including: receiving one or more observation data related to a predicted event, and analyzing the one or more observation data based on an event prediction model and outputting an event level at which the predicted event occurs, wherein the event prediction model includes a hierarchical analysis network, the hierarchical analysis network includes multiple types of prior events, each type of prior event includes one or more representative events, and each type of prior event corresponds to an event level at which the predicted event occurs.

[0005] In a second aspect of the present disclosure, there is provided an information processing device for event prediction, including a processor configured to execute the above event prediction method.

[0006] The event prediction method and information processing device provided by the embodiments of the present invention use an event prediction model constructed by a hierarchical analysis network to predict which event level the occurrence of an event belongs to by systematically analyzing observation data layer by layer and step by step through several types of prior events, the representative events corresponding to each type of prior event, and multiple event levels corresponding to each type of prior event, thereby improving the prediction accuracy and breadth of events.

[0007] It should be understood that the content described in the part of the summary of the invention is not intended to limit the key features or important features of the embodiments of the present disclosure, nor is it used to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. Brief Description of the Drawings

[0008] In combination with the drawings and referring to the following detailed description, the above and other features, advantages, and aspects of the embodiments of the present disclosure will become more obvious. Among them:

[0009] Figure 1 A flowchart showing a method for event prediction in an embodiment of the present disclosure is shown.

[0010] Figure 2 The figure shows a schematic diagram of the model architecture of an event prediction model adopted by the event prediction method in an embodiment of the present disclosure. Detailed implementation manners

[0011] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not used to limit the protection scope of the present disclosure.

[0012] Please refer to Figure 1 , an embodiment of the present invention provides an event prediction method 10, and the method 10 includes:

[0013] S12, receiving one or more observation data related to the predicted event, and

[0014] S14, analyzing the one or more observation data based on an event prediction model and outputting an event level at which the predicted event occurs, where the event prediction model includes a hierarchical analysis network, the hierarchical analysis network includes multiple types of prior events, each type of prior event includes one or more characterizing events, and each type of prior event corresponds to an event level at which the predicted event occurs.

[0015] The event prediction method 10 can be applied to multiple technical fields, such as power, network fault prediction, energy distribution prediction, production decision-making, transportation, environmental monitoring, and military, etc. The event prediction can be fault prediction, natural disaster risk assessment, energy distribution estimation, etc.

[0016] In the above step S12, the one or more observation data can be obtained according to the application scenario of the predicted event, and the one or more observation data may include one or more actually occurring observation events, specific information or data of the observation events, and time information of the occurrence of the observation events. The one or more observation data can be obtained by an information collection device according to the application scenario. The information collection device can be, but is not limited to, one or more of sensors, detectors, and network communication devices. The one or more observation data can reflect the specific situation of the observation events so as to predict the possibility of future events based on the specific situation.

[0017] In the step S14, the event prediction model is used to determine the event level at which the predicted event occurs by using the input one or more observation data.

[0018] Please refer to Figure 1 and Figure 2 simultaneously. The event prediction method 10 further includes: step S11 of constructing the event prediction model.

[0019] The event prediction model includes an analytic hierarchy process network constructed by the analytic hierarchy process method. Step S11 of constructing the event prediction model further includes:

[0020] S112. For each prediction event, construct multiple event levels;

[0021] S114. For each event level, construct multiple types of prior events;

[0022] S116, and for each type of prior event, construct multiple representative events, and

[0023] S118. Determine the probability of occurrence of the prediction event based on the prediction probabilities of the representative events, prior events, and event levels.

[0024] In an embodiment of the present disclosure, in the event prediction model, multiple event levels X i can be constructed for each prediction event, for example, three event levels: primary, obvious, and immediate before. The prediction probabilities of each event level can be expressed as X1, X2, X3, and their value ranges are respectively set as [0, s i , i = 1, 2, 3. Where s i can be set to different values according to different event levels. The three event levels of primary, obvious, and immediate before represent different degrees and possibilities of the occurrence of the prediction event. Primary means that when the corresponding representative event occurs, it may cause the occurrence of the prediction event. Obvious means that when the corresponding representative event occurs, it is very likely to cause the occurrence of the prediction event. Immediate before means that when the corresponding representative event occurs, it can basically be determined that the prediction event occurs. It can be understood that the event levels described in the embodiments of the present disclosure are not limited to three.

[0025] In an embodiment of the present disclosure, the event level X i includes m types of prior events Y ij related to the prediction event, 1 ≤ j ≤ m. The m types of prior events Y ij can be obtained through comprehensive analysis of expert knowledge or historical data. The prediction probability value range of the prior event Y ij is [0, q ij .

[0026] In an embodiment of the present disclosure, the prior event Y ij may include n specific representative events Z ijk, where 1 ≤ k ≤ n, the prediction method 100 can further set a basic prediction probability characterizing event Z according to historical data ijk of the time effect parameter and the time effect decay function and its parameters.

[0027] In an embodiment of the present disclosure, constructing the event prediction model can further configure a parameter reflecting whether the predicted event has an influence at a certain moment. Specifically, the parameter can be a constant ε. When calculating the prediction probability P ijk at time T ijk of the characterizing event Z ijk (T), if P ijk (T) < ε, it is considered that the characterizing event Z ijk has no influence at time T. In other words, this characterizing event Z ijk has no effect on the prediction of the predicted event at time T.

[0028] In an embodiment of the present disclosure, the step S14 further includes:

[0029] S142, determining one or more first characterizing events associated with the one or more observation data based on the one or more observation data;

[0030] S144, determining one or more first prior events corresponding to the first characterizing events, and

[0031] S146, determining the event level of the occurrence of the predicted event based on the one or more first prior events.

[0032] In an embodiment of the present disclosure, the step S142 further includes:

[0033] S1422, determining the prediction probability of the one or more characterizing events at the prediction time based on the one or more observation data, and

[0034] S1424, determining one or more of the first characterizing events based on the prediction probability.

[0035] In an embodiment of the present disclosure, the prediction probability P ijk of the characterizing event Z corresponding to the observation data at the current time T, the event occurrence time T ijk , the time effect decay function and its parameters can be used to calculate the prediction probability P ijk of the characterizing event Z at time T ijk(T). The time effect attenuation function may be, but is not limited to, a Gaussian function. The prediction probability of one or more of the characterization events at the prediction moment is determined based on the time effect attenuation function:

[0036]

[0037] The parameter is the basic prediction probability of the characterization event Z ijk , the parameter T ijk is the time when the characterization event Z ijk occurs, the parameter is the time effect parameter, the parameters A and σ are the parameters of the time attenuation function. If P ijk (T) < ε, it means that the characterization event Z ijk has no influence at time T and subsequent prediction of event levels may not be used. The prediction probabilities of one or more of the characterization events at the prediction moment can be arranged from large to small for subsequent analysis of prior events.

[0038] In an embodiment of the present disclosure, the step S144 further includes:

[0039] S1442, determining the prediction probability of each type of prior event at the prediction moment based on the prediction probabilities of one or more first characterization events at the prediction moment, and

[0040] S1444, determining one or more of the first prior events based on the prediction probabilities of each type of prior event at the prediction moment.

[0041] In an embodiment of the present disclosure, the prediction probability of the prior event at the prediction moment is obtained by the following calculation method:

[0042]

[0043] The P' ijk (T) is the prediction probability of the prior event Y ij at the prediction moment T. The value range of the prediction probability of the prior event Y ij is [0, q ij . If all the characterization events in the prior event Y ij have influence at time T, the prediction probability of Yij takes its upper limit q ij , otherwise the prediction probability of the prior event Y ij at time T is calculated by the above P' ijk (T) formula. The formula P' ijk (T) is iterated n times, and P' ijn (T) is the prediction probability of the prior event Y ijThe predicted probability P' ij , where the parameter n is the number of the characterization events corresponding to the prior event Y ij among them. The calculated multiple influential predicted probabilities P' ij can be arranged in descending order according to the probability values for subsequent calculation of the event level X i of the predicted probability.

[0044] In an embodiment of the present disclosure, the step S146 further includes:

[0045] S1462, determining the predicted probability of each event level at which one or more of the first prior events occur based on the predicted probability of one or more of the first prior events at the prediction moment, and

[0046] S1464, determining the probability of the event level at which the predicted event occurs according to the magnitudes of the predicted probabilities of each event level.

[0047] In an embodiment of the present disclosure, the predicted probability of the event level is obtained by the following calculation method:

[0048]

[0049] where the predicted probability of the event level X i is P'' i (T), the range of the predicted probability P'' i (T) is [0, s i , after m iterations, P″ im (T) is the predicted probability of the event level X at time T i , denoted as P″ i , and the parameter m represents the number of the prior events.

[0050] The probability P(T) of the event level at which the predicted event occurs is:

[0051] P(T) = max(P″ i1 , P″ i2 , P″ i3 …),

[0052] The parameters i1, i2, i3 represent the numbers of the event levels.

[0053] In an embodiment of the present disclosure, the event level of the predicted event can be determined by obtaining the probability of the event level at which the predicted event occurs, and the probability of the event to be predicted can be determined according to the level of the predicted event.

[0054] In an embodiment of the present disclosure, the event prediction method 10 further includes obtaining an event level at which the predicted event occurs at a specific moment.

[0055] In an embodiment of the present disclosure, the event prediction method 10 further includes generating one or more policy information based on the level at which the predicted event occurs.

[0056] The event prediction method provided by the embodiments of the present disclosure is an event prediction method based on trend logic. By using an event prediction model constructed by an analytic hierarchy network, the possibility of an event occurring can be predicted through a systematic and hierarchical analysis combining qualitative and quantitative methods. Specifically, by systematically analyzing the observed data layer by layer and step by step through several types of prior events, the representative events corresponding to each type of prior event, and multiple event levels corresponding to each type of prior event, it can be predicted which event level the event belongs to when it occurs, thereby improving the prediction ability and accuracy of the event.

[0057] An embodiment of the present invention further provides an information processing device for the event prediction. The information processing device at least includes a processor that can execute all steps in the event prediction method 10 described above, which will not be elaborated here again. The information processing device can be any computer or other electronic device with data analysis capabilities. The information processing device may further include a computer storage medium for storing the observed data and the event prediction model.

[0058] The medium can be any accessible medium that can be obtained by the information processing device, including but not limited to volatile and non-volatile media, removable and non-removable media. The memory can be a volatile memory (such as registers, caches, random access memory (RAM)), a non-volatile memory (such as read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory), or some combination thereof. The storage device can be a removable or non-removable medium and can include machine-readable media, such as a flash drive, a magnetic disk, or any other medium.

[0059] Aspects of the present disclosure are described herein with reference to the flowcharts and / or block diagrams of methods, apparatuses, devices, and computer program products implemented according to the present disclosure. It should be understood that each block of the flowcharts and / or block diagrams, and the combinations of blocks in the flowcharts and / or block diagrams, can be implemented by computer-readable program instructions.

[0060] Computer-readable program instructions may be loaded onto a computer, other programmable data processing apparatus, or other device, causing a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to generate a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions / acts specified in one or more blocks of the flowchart and / or block diagram.

[0061] The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various implementations of the present disclosure. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a portion of an instruction, which contains one or more executable instructions for implementing the specified logical function. In some alternative implementations, the functions noted in the blocks may occur out of the order noted in the figures. For example, two consecutive blocks may in fact be executed substantially in parallel, or they may sometimes be executed in the reverse order, depending upon the functionality involved. It should also be noted that each block of the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented by special purpose hardware-based systems that perform the specified functions or acts, or by combinations of special purpose hardware and computer instructions.

[0062] The various implementations of the present disclosure have been described above. The above description is exemplary, not exhaustive, and is not limited to the disclosed implementations. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described implementations. The choice of terms used herein is intended to best explain the principles of the implementations, the practical application, or the improvement of the technology in the marketplace, or to enable other ordinary skill in the art to understand the implementations disclosed herein.

Claims

1. A method for event prediction, characterized in that, Comprising: Receiving one or more pieces of observation data related to a predicted event, and Analyzing the one or more pieces of observation data based on an event prediction model and outputting an event level of the occurrence of the predicted event, wherein the event prediction model includes an analytic hierarchy network, the analytic hierarchy network includes multiple types of prior events, each type of prior event includes one or more characterizing events, and each type of prior event corresponds to an event level of the occurrence of the predicted event; Further analyzing the one or more pieces of observation data based on the event prediction model includes: determining one or more first characterizing events based on the one or more pieces of observation data; determining one or more first prior events corresponding to the first characterizing events, and determining the event level of the occurrence of the predicted event based on the one or more first prior events; Determining one or more of the first characterizing events further includes: determining the prediction probabilities of the one or more characterizing events at a prediction moment based on the one or more pieces of observation data, and determining one or more of the first characterizing events based on the prediction probabilities; The prediction probabilities of the one or more characterizing events at the prediction moment are determined based on a time effect decay function: Among them, the characterization event Z ijk The predicted probability at the predicted time T is denoted as P ijk (T), where parameter i corresponds to the level, j corresponds to the j-th prior event, k corresponds to the k-th said characterization event, and the parameter is the basic predicted probability of the characterization event Z ijk , parameter T ijk is the time when the characterization event Z ijk occurs, parameter is the time effect parameter, parameters A and σ are the parameters of the time effect attenuation function. If P ijk (T) < ε, it means that the characterization event Z ijk has no influence at time T, and the parameter ε is a constant.

2. The event prediction method according to any one of claims 1, characterized in that Determining one or more of the first prior events further includes: Determining the prediction probabilities of each type of prior event at the prediction moment based on the prediction probabilities of the one or more first characterizing events at the prediction moment, and Determining one or more of the first prior events based on the prediction probabilities of each type of prior event at the prediction moment.

3. The event prediction method according to claim 2, wherein The prediction probabilities of the prior events at the prediction moment are obtained through the following calculation method: The said P' ijk (T) is the prior event Y ij at the prediction probability at the prediction moment T, and the prior event Y ij has a value range of [0, q ij . If all the representation events in the prior event Y ij are influential at the moment T, then the prediction probability of Y ij takes its upper limit q ij . Otherwise, the prediction probability of the prior event Y ij at the moment T is calculated by the above P' ijk (T) formula. The formula P' ijk (T) has been iterated n times, and P' ijn (T) is the prediction probability P' ij of the prior event Y ij at the moment T. The parameter n is the number of the corresponding representation events in the prior event Y ij .

4. The event prediction method according to claim 3, wherein Further determining the event level of the occurrence of the predicted event based on the one or more first prior events includes: Determining the prediction probabilities of each event level of the occurrence of the one or more first prior events based on the prediction probabilities of the one or more first prior events at the prediction moment, and Determining the probability of the event level of the occurrence of the predicted event according to the magnitudes of the prediction probabilities of each event level.

5. The event prediction method according to claim 4, wherein The prediction probabilities of the event levels are obtained through the following calculation method: where the event level is X i The predicted probability is P” i (T), and the range of the predicted probability P” i (T) is [0, s i , after m iterations, P” im (T) is the predicted probability of the event level X at time T i and is denoted as P” i , where the parameter m represents the number of prior events; The probability P(T) of the event level of the occurrence of the predicted event is: P(T) = max(P” i1 , P” i2 , P” i3 …) The parameters i1, i2, i3 represent the numbers of the event levels.

6. The event prediction method according to claim 1, characterized in that The one or more pieces of observation data include one or more observation events and time information of the occurrence of the one or more observation events.

7. The event prediction method according to claim 1, wherein Further including obtaining the event level of the occurrence of the predicted event at a specific moment.

8. The event prediction method according to claim 1, wherein Further including: Generating one or more pieces of policy information based on the level of the occurrence of the predicted event.

9. The event prediction method according to claim 1, wherein, The event levels include three levels: primary, obvious, and immediate before.

10. An information processing device for event prediction, comprising a processor, characterized in that, The processor is configured to execute the event prediction method according to any one of claims 1-9.

Images