Event early warning method, device and equipment based on big data analysis, and storage medium

Abstract

The application belongs to the technical field of event early warning, and discloses an event early warning method, device and equipment based on big data analysis and a storage medium. The method comprises the following steps: when an abnormal signal of a multi-source event is monitored, determining an abnormal event in the multi-source event according to the abnormal signal; obtaining historical data of the abnormal event in a historical time period, and determining a prediction model based on the historical data; inputting the abnormal event into the prediction model to determine a prediction result of the abnormal event; judging whether early warning of the abnormal event is needed according to the prediction result; when it is determined that early warning of the abnormal event is needed, inputting the abnormal event and the prediction result into a hierarchical model to determine an early warning level of the abnormal event, and completing early warning of the abnormal event according to the early warning level. The multi-source data can be automatically analyzed and early warned.

Description

Technical Field

[0001] This invention relates to the field of event early warning technology, and in particular to an event early warning method, device, equipment and storage medium based on big data analysis. Background Technology

[0002] Currently, the analysis and early warning of multi-source events mainly rely on manual work. However, manual analysis and early warning are inefficient and fail to identify event risks in advance, which may lead to huge financial losses for the company. Summary of the Invention

[0003] The main objective of this invention is to provide an event early warning method, device, equipment, and storage medium based on big data analysis, aiming to solve the technical problem in the prior art where the inefficiency of manual analysis and early warning methods leads to the inability to identify event risks in advance.

[0004] To achieve the above objectives, the present invention provides an event early warning method based on big data analysis, the method comprising the following steps:

[0005] When an abnormal signal is detected in a multi-source event, the abnormal event in the multi-source event is determined based on the abnormal signal;

[0006] Obtain historical data of the abnormal event within a historical time period, and determine a prediction model based on the historical data;

[0007] The abnormal event is input into the prediction model to determine the prediction result of the abnormal event;

[0008] Determine whether an early warning is needed for the abnormal event based on the prediction results;

[0009] When it is determined that an early warning is needed for the abnormal event, the abnormal event and the prediction result are input into the hierarchical model to determine the early warning level of the abnormal event, and the early warning of the abnormal event is completed according to the early warning level.

[0010] Optionally, before determining the abnormal event among the multi-source events based on the abnormal signal when an abnormal signal is detected in the multi-source events, the method further includes:

[0011] Obtain high-frequency and low-frequency data for each event in the multi-source event;

[0012] By correlating the high-frequency data with the monitoring frequency of the low-frequency data, key information can be obtained;

[0013] An initial monitoring model is constructed based on the high-frequency data;

[0014] The initial monitoring model is estimated using the key information to obtain the monitoring model;

[0015] The multi-source events are input into the monitoring model to obtain monitoring results, wherein the monitoring results are used to determine whether abnormal signals occur in the multi-source events.

[0016] Optionally, obtaining historical data of the abnormal event within a historical time period and determining a prediction model based on the historical data includes:

[0017] Obtain historical data of the aforementioned abnormal event within the historical time period;

[0018] Based on the historical data, determine the duration of the abnormal event within the historical time period, and filter out the target duration period that is greater than the target threshold from the duration period.

[0019] Determine the feature vector of the abnormal event based on the target duration segment;

[0020] The feature vector is input into the initial prediction model to obtain the prediction model.

[0021] Optionally, the step of inputting the abnormal event and the prediction result into the hierarchical model to determine the warning level of the abnormal event includes:

[0022] Based on the historical data, determine the first probability that the abnormal event will trigger an early warning in the current time period;

[0023] Based on the historical data, a second probability of the abnormal event triggering an early warning within different periods is determined;

[0024] Establish a relationship curve between the historical data and the average number of warnings for the abnormal event within a historical time period, and determine a third possibility based on the relationship curve;

[0025] Based on the first possibility, the second possibility, and the third possibility, an initial prediction model is trained to obtain the hierarchical model;

[0026] The abnormal event and the prediction result are input into the hierarchical model to determine the warning level of the abnormal event.

[0027] Optionally, the warning levels include a first warning level, a second warning level, and a third warning level; wherein,

[0028] The warning for the abnormal event is completed according to the warning level, including:

[0029] If the warning level is the first warning level, the warning of the abnormal event is completed by playing a voice message.

[0030] If the warning level is the second warning level, the warning of the abnormal event is completed through visual display;

[0031] If the warning level is the third warning level, the abnormal event is interpreted to obtain the interpretation result, and the interpretation result is displayed visually to complete the warning of the abnormal event.

[0032] Optionally, the step of interpreting the abnormal event to obtain an interpretation result, and then visually displaying the interpretation result to complete the early warning of the abnormal event, includes:

[0033] Statistical analysis and machine learning are used to identify patterns, trends, and outliers in the abnormal events, and the patterns, trends, and outliers are used as the interpretation results of the abnormal events.

[0034] The visualization format is determined, and the interpretation results are displayed visually based on the visualization format to complete the early warning of the abnormal event. The visualization format includes charts, graphs, and images.

[0035] Optionally, the event early warning method based on big data analysis further includes:

[0036] After issuing an early warning for the abnormal event based on the warning level, the number of warnings for the abnormal event within a future time period is obtained.

[0037] A threshold number is set, and when the number of warnings exceeds the threshold number, a rapid warning mode for the abnormal event is activated;

[0038] When an abnormal event is detected as having an abnormal signal in the multi-source event while the abnormal event is in the rapid early warning mode, an early warning is issued.

[0039] Furthermore, to achieve the above objectives, the present invention also proposes an event early warning device based on big data analysis, wherein the event early warning device based on big data analysis includes:

[0040] The determination module is used to determine the abnormal event among the multi-source events based on the abnormal signal when an abnormal signal is detected in the multi-source events.

[0041] The acquisition module is used to acquire historical data of the abnormal event within a historical time period and determine a prediction model based on the historical data.

[0042] The determining module is used to input the abnormal event into the prediction model and determine the prediction result of the abnormal event;

[0043] The judgment module is used to determine whether an early warning is needed for the abnormal event based on the prediction result.

[0044] The early warning module is used to input the abnormal event and the prediction result into the hierarchical model when it is determined that an early warning for the abnormal event is required, to determine the early warning level of the abnormal event, and to complete the early warning of the abnormal event according to the early warning level.

[0045] Furthermore, to achieve the above objectives, the present invention also proposes an event early warning device based on big data analysis. The event early warning device based on big data analysis includes: a memory, a processor, and an event early warning program based on big data analysis stored in the memory and executable on the processor. The event early warning program based on big data analysis is configured to implement the steps of the event early warning method based on big data analysis as described above.

[0046] Furthermore, to achieve the above objectives, the present invention also proposes a storage medium storing an event warning program based on big data analysis, wherein when the event warning program based on big data analysis is executed by a processor, it implements the steps of the event warning method based on big data analysis as described above.

[0047] This invention proposes an event early warning method, apparatus, device, and storage medium based on big data analysis. When abnormal signals are detected in multi-source events, the method identifies the abnormal events within those events based on the abnormal signals; acquires historical data of the abnormal events within a historical time period and determines a prediction model based on the historical data; inputs the abnormal events into the prediction model to determine the prediction result; determines whether an early warning is needed based on the prediction result; and, if an early warning is needed, inputs the abnormal events and the prediction result into a hierarchical model to determine the early warning level of the abnormal events and completes the early warning based on the warning level. Through this method, it is possible to monitor for abnormal signals in multi-source events in real time, identify abnormal events based on these signals, and then use a prediction model to predict whether an early warning is needed. This achieves automatic analysis and early warning of multi-source data, and improves the efficiency of analysis and early warning while identifying event risks in advance, thereby effectively preventing significant losses to the company due to abnormal events. Attached Figure Description

[0048] Figure 1 This is a schematic diagram of the structure of an event early warning device based on big data analysis in the hardware operating environment involved in the embodiments of the present invention;

[0049] Figure 2This is a flowchart illustrating the first embodiment of the event early warning method based on big data analysis of the present invention.

[0050] Figure 3 This is a flowchart illustrating the second embodiment of the event early warning method based on big data analysis of the present invention.

[0051] Figure 4 This is a structural block diagram of the first embodiment of the event early warning device based on big data analysis of the present invention.

[0052] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0053] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0054] Reference Figure 1 , Figure 1 This is a schematic diagram of the structure of an event early warning device based on big data analysis in the hardware operating environment of the embodiment of the present invention.

[0055] like Figure 1 As shown, the event early warning device based on big data analysis may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen and an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wireless-Fidelity (Wi-Fi) interface). The memory 1005 may be a high-speed random access memory (RAM) or a stable non-volatile memory (NVM), such as a disk drive. The memory 1005 may also optionally be a storage device independent of the aforementioned processor 1001.

[0056] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on event early warning devices based on big data analytics, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0057] like Figure 1As shown, the memory 1005, which serves as a storage medium, may include an operating system, a network communication module, a user interface module, and an event warning program based on big data analysis.

[0058] exist Figure 1 In the event early warning device based on big data analysis shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the event early warning device based on big data analysis of the present invention can be set in the event early warning device based on big data analysis. The event early warning device based on big data analysis calls the event early warning program based on big data analysis stored in the memory 1005 through the processor 1001 and executes the event early warning method based on big data analysis provided in the embodiment of the present invention.

[0059] Based on the above hardware structure, an embodiment of the event early warning method based on big data analysis of the present invention is proposed.

[0060] Reference Figure 2 , Figure 2 This is a flowchart illustrating the first embodiment of an event early warning method based on big data analysis according to the present invention.

[0061] In this embodiment, the event early warning method based on big data analysis includes the following steps:

[0062] Step S10: When an abnormal signal is detected in a multi-source event, the abnormal event in the multi-source event is determined based on the abnormal signal.

[0063] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a mobile phone, tablet computer, or personal computer, or an electronic device capable of performing the above functions or an event early warning device based on big data analysis. The following description uses the event early warning device based on big data analysis as an example to illustrate this embodiment and the subsequent embodiments.

[0064] It should be noted that multi-source events include multiple events, which can be internal financial events, economic events, online public opinion events, etc.; abnormal events refer to events in multi-source events that show abnormal signals.

[0065] In practical implementation, an abnormal signal in a multi-source event can be an abnormal signal from one of the events in the multi-source event, or it can be an abnormal signal from multiple events in the multi-source event. Specifically, taking an abnormal signal from an online public opinion event in a multi-source event as an example, a preset number of occurrences can be set. When the number of times online public opinion about the company is obtained within the preset event segment is greater than the preset number of occurrences, it can be determined that an abnormal signal has occurred in the online public opinion event in the multi-source event.

[0066] In practical implementation, corresponding thresholds can be set for all events in multi-source events, and the number of times the target scene appears in the event within a preset time period can be determined. Then, the number of occurrences is compared with the corresponding threshold. When the number of occurrences is greater than the corresponding threshold, it can be considered that an abnormal signal has been detected in the multi-source event.

[0067] In one embodiment, before determining the abnormal event among the multi-source events based on the abnormal signal when an abnormal signal is detected in the multi-source events, the method further includes:

[0068] Obtain high-frequency and low-frequency data for each event in the multi-source event;

[0069] By correlating the high-frequency data with the monitoring frequency of the low-frequency data, key information can be obtained;

[0070] An initial monitoring model is constructed based on the high-frequency data;

[0071] The initial monitoring model is estimated using the key information to obtain the monitoring model;

[0072] The multi-source events are input into the monitoring model to obtain monitoring results, wherein the monitoring results are used to determine whether abnormal signals occur in the multi-source events.

[0073] In practice, high-frequency data can be daily data of events, while low-frequency data can be monthly or annual data of events. Correlating high-frequency data with low-frequency data monitoring frequencies enables the effective utilization of key information at different data frequencies.

[0074] It is understandable that since the key information is obtained by associating high-frequency data with low-frequency data, if there is no corresponding observation in the key information, it is recorded as a missing value. When estimating the initial monitoring model, the Kalman filtering method can be used to ensure that the model estimation is not affected by individual missing values, so that the monitoring model can be more accurate when monitoring abnormal signals in multi-source events.

[0075] It should be noted that the initial monitoring model is an untrained monitoring model, while the monitoring model is a trained model that can be directly used to monitor whether abnormal signals occur in multi-source events; the monitoring model can be one or more of the following: a convolutional neural network model or a deep neural network model.

[0076] In this embodiment, a monitoring model is used to monitor in real time whether abnormal signals appear in multi-source events. When an abnormal signal is detected in a multi-source event, the abnormal event in the multi-source event is determined based on the abnormal signal. This can initially screen out abnormal events that may require early warning from the multi-source events, preventing the direct judgment of whether a multi-source event needs early warning when there are too many events, which would affect the efficiency of analysis and early warning, thereby effectively improving the efficiency of analysis and early warning.

[0077] Step S20: Obtain historical data of the abnormal event within a historical time period, and determine a prediction model based on the historical data.

[0078] In practice, the prediction model can be constructed using any of the following: feedforward neural networks, convolutional neural networks, recurrent neural networks, or long short-term memory networks.

[0079] In one embodiment, obtaining historical data of the abnormal event within a historical time period and determining a prediction model based on the historical data includes:

[0080] Obtain historical data of the aforementioned abnormal event within the historical time period;

[0081] Based on the historical data, determine the duration of the abnormal event within the historical time period, and filter out the target duration period that is greater than the target threshold from the duration period.

[0082] Determine the feature vector of the abnormal event based on the target duration segment;

[0083] The feature vector is input into the initial prediction model to obtain the prediction model.

[0084] It should be noted that the target threshold can be set in advance. Segments with a duration greater than the target threshold can be considered valid segments, while segments with a duration less than or equal to the target threshold can be considered invalid segments. The target segment is the valid segment. Since the invalid segment has a duration less than the target threshold, it can be considered that the invalid segment cannot reflect the relevant characteristics when the abnormal event occurs. Therefore, it is necessary to remove the segments with a duration less than or equal to the target threshold, and then use only the data corresponding to the target segment to train the initial prediction model. This will make the prediction results of the prediction model more accurate.

[0085] It is understandable that by inputting the feature vector into the initial prediction model, the output results of the abnormal events can be obtained. Based on the label of the abnormal events and the output results, the loss value is determined. Then, by continuously updating the parameters in the initial prediction model according to the loss value, the prediction model can be obtained.

[0086] In this embodiment, the duration of the abnormal event within a historical time period is first determined by using historical data. Then, target duration segments with a duration greater than the target threshold are selected from the duration segments. Finally, the feature vector of the abnormal event is determined based on the target duration segment to train the initial prediction model, which can make the prediction accuracy of the trained prediction model higher.

[0087] Step S30: Input the abnormal event into the prediction model and determine the prediction result of the abnormal event.

[0088] It should be noted that the forecast results include those that require an early warning and those that do not.

[0089] Step S40: Determine whether an early warning is needed for the abnormal event based on the prediction results.

[0090] It should be noted that the prediction results may include those that require an early warning and those that do not.

[0091] Step S50: When it is determined that an early warning is needed for the abnormal event, the abnormal event and the prediction result are input into the hierarchical model to determine the early warning level of the abnormal event, and the early warning of the abnormal event is completed according to the early warning level.

[0092] It should be noted that when an abnormal event is determined to require an early warning, the abnormal event and the prediction result can be directly input into the hierarchical model to determine the warning level of the abnormal event.

[0093] Understandably, by first monitoring whether abnormal events show abnormal signals, and then determining whether an early warning is needed for the abnormal event, it is possible to effectively prevent the occurrence of early warning errors.

[0094] In practice, different warning levels can correspond to different protective measures, and different warning levels can also correspond to different warning forms. Users can take corresponding protective measures based on the warning level of the abnormal event, and the system can also complete the warning of abnormal events based on the warning form corresponding to the warning level.

[0095] This embodiment detects abnormal signals in multi-source events, identifies abnormal events within those events based on these signals, acquires historical data of the abnormal events over a given time period, and determines a prediction model based on this data. The abnormal event is then input into the prediction model to determine the prediction result. Based on the prediction result, it is determined whether an early warning is needed. If an early warning is required, the abnormal event and the prediction result are input into a hierarchical model to determine the warning level, and an early warning is issued based on this level. This method enables real-time monitoring of abnormal signals in multi-source events, identification of abnormal events based on these signals, and prediction of these events using a prediction model to determine the need for an early warning. This allows for automatic analysis and early warning of multi-source data, improving analysis and warning efficiency while proactively identifying event risks, thus effectively preventing significant losses to the company due to abnormal events.

[0096] refer to Figure 3 , Figure 3 This is a flowchart illustrating a second embodiment of an event early warning method based on big data analysis according to the present invention.

[0097] Based on the first embodiment described above, the event early warning method based on big data analysis in this embodiment inputs the abnormal event and the prediction result into a hierarchical model to determine the early warning level of the abnormal event, including:

[0098] Step S301: Determine the first probability that the abnormal event will trigger an early warning in the current time period based on the historical data.

[0099] It should be noted that the current time period refers to the time from when the abnormal signal of the abnormal event was detected to the present moment; the first probability refers to the probability that the abnormal event will trigger an early warning during the current time period.

[0100] Step S302: Determine a second probability that the abnormal event will trigger an early warning in different periods based on the historical data.

[0101] It should be noted that different cycles refer to different cycles of the abnormal event within a historical time period, such as a month within a historical time period, or a week within a historical time period, etc.; the second probability refers to the probability of the abnormal event issuing a warning within different cycles.

[0102] Step S303: Establish a relationship curve between the historical data and the average number of warnings for the abnormal event within a historical time period, and determine a third possibility based on the relationship curve.

[0103] It should be noted that the relationship curve can be used to represent the average number of warnings in historical data over various historical time periods. The relationship curve can be compared with the preset number of warnings to determine the third possibility. The third possibility refers to the ratio of the number of historical time periods in the relationship curve where the average number of warnings is greater than the preset number of warnings to the total number of historical time periods in the relationship curve. Specifically, for example, if the average number of warnings for an abnormal event on the first day in the past is 1, the average number of warnings on the second day in the past is 2, and the average number of warnings on the third day in the past is 1, and the preset number of warnings is set to 1, then the total number of historical time periods in the relationship curve can be determined to be 3, and the number of historical time periods in the relationship curve where the average number of warnings is greater than the preset number of warnings is 1. Finally, the third possibility can be obtained as 1 / 3. The preset number of warnings can be set in advance.

[0104] Step S304: Train an initial prediction model based on the first possibility, the second possibility, and the third possibility to obtain the hierarchical model.

[0105] Step S305: Input the abnormal event and the prediction result into the hierarchical model to determine the warning level of the abnormal event.

[0106] It should be noted that the hierarchical model can be constructed using classification network models, such as VGG, Googlenet, ResNet50, and MobileNetV2, etc.

[0107] Understandably, it is possible to pre-set warning levels and corresponding warning formats for different warning levels, and then issue warnings based on the warning formats.

[0108] In one embodiment, the warning levels include a first warning level, a second warning level, and a third warning level; wherein,

[0109] The warning for the abnormal event is completed according to the warning level, including:

[0110] If the warning level is the first warning level, the warning of the abnormal event is completed by playing a voice message.

[0111] If the warning level is the second warning level, the warning of the abnormal event is completed through visual display;

[0112] If the warning level is the third warning level, the abnormal event is interpreted to obtain the interpretation result, and the interpretation result is displayed visually to complete the warning of the abnormal event.

[0113] It should be noted that the third warning level is more urgent than the second warning level, and the second warning level is more urgent than the first warning level.

[0114] In practice, when the warning level is the first warning level, it indicates that the urgency of the abnormal event is relatively low, and the abnormal event can be broadcast via voice. When the warning level is the second warning level, the abnormal event needs to be displayed visually. When the warning level is the third warning level, it indicates that the urgency of the abnormal event is relatively high, and the abnormal event needs to be interpreted, and then the interpretation results should be displayed visually. This allows relevant staff within the company to quickly understand the abnormal information and data, and then take timely measures to prevent the abnormal event from causing huge losses to the company.

[0115] In one embodiment, the step of interpreting the abnormal event to obtain an interpretation result and then visually displaying the interpretation result to complete the early warning of the abnormal event includes:

[0116] Statistical analysis and machine learning are used to identify patterns, trends, and outliers in the abnormal events, and the patterns, trends, and outliers are used as the interpretation results of the abnormal events.

[0117] The visualization format is determined, and the interpretation results are displayed visually based on the visualization format to complete the early warning of the abnormal event. The visualization format includes charts, graphs, and images.

[0118] It should be noted that visualizations can include charts, graphs, images, etc., and can make the data easier to understand and interpret through color, arrangement, scaling, and other methods.

[0119] In practical implementation, large-screen optimization technology needs to be used when visualizing the display. For example, the layout of the visualization can be adjusted to adapt to the size and resolution of the large screen, ensuring the clarity and readability of the information. The visual effect of the information display and the user experience can also be improved by adjusting the font and icon size and designing animation effects.

[0120] In this embodiment, by interpreting abnormal events and then visualizing the interpretation results according to the visualization format selected by the user, complex data can be displayed in an intuitive and easy-to-understand way, providing a good user experience.

[0121] In one embodiment, the event early warning method based on big data analysis further includes:

[0122] After issuing an early warning for the abnormal event based on the warning level, the number of warnings for the abnormal event within a future time period is obtained.

[0123] A threshold number is set, and when the number of warnings exceeds the threshold number, a rapid warning mode for the abnormal event is activated;

[0124] When an abnormal event is detected as having an abnormal signal in the multi-source event while the abnormal event is in the rapid early warning mode, an early warning is issued.

[0125] It should be noted that the future time period can be preset, such as the next week, the next month, or the next month, etc.; the rapid warning mode means that an alert can be issued as soon as an abnormal signal of an event is detected, without the need for the prediction module to predict whether an alert is needed.

[0126] In the specific implementation, after the abnormal event warning is completed, it is necessary to continuously monitor the number of abnormal event warnings in the future. When the number of abnormal event warnings exceeds the threshold, a fast warning mode can be activated. After the fast warning mode is activated, a warning can be issued directly when an abnormal event signal is detected again.

[0127] In this embodiment, when multiple warnings of abnormal events occur within a future time period, a warning can be issued immediately after the abnormal signal of the abnormal event is detected, which can effectively improve the warning efficiency and prevent the company from suffering significant losses due to untimely warnings.

[0128] This embodiment determines the first probability of an abnormal event triggering an alert in the current time period based on historical data; determines the second probability of the abnormal event triggering an alert in different periods based on historical data; establishes a relationship curve between the historical data and the average number of alerts for the abnormal event over a historical time period, and determines a third probability based on the relationship curve; trains an initial prediction model based on the first, second, and third probabilities to obtain the hierarchical model; and inputs the abnormal event and the prediction results into the hierarchical model to determine the alert level of the abnormal event. Through this method, the alert level of an abnormal event can be quickly and accurately determined using a hierarchical model, enabling company employees to take timely and appropriate measures to rectify the abnormal event based on different alert levels, preventing significant losses to the company.

[0129] Furthermore, this embodiment of the invention also proposes a storage medium storing an event warning program based on big data analysis. When the event warning program based on big data analysis is executed by a processor, it implements the steps of the event warning method based on big data analysis as described above.

[0130] Reference Figure 4 , Figure 4 This is a structural block diagram of the first embodiment of the event early warning device based on big data analysis of the present invention.

[0131] like Figure 4 As shown, the event early warning device based on big data analysis proposed in this embodiment of the invention includes:

[0132] The determination module 10 is used to determine the abnormal event among the multi-source events based on the abnormal signal when an abnormal signal is detected in the multi-source events.

[0133] The acquisition module 20 is used to acquire historical data of the abnormal event within a historical time period and determine a prediction model based on the historical data.

[0134] The determining module 10 is used to input the abnormal event into the prediction model and determine the prediction result of the abnormal event.

[0135] The judgment module 30 is used to determine whether an early warning is needed for the abnormal event based on the prediction result.

[0136] The early warning module 40 is used to input the abnormal event and the prediction result into the hierarchical model when it is determined that an early warning of the abnormal event is required, to determine the early warning level of the abnormal event, and to complete the early warning of the abnormal event according to the early warning level.

[0137] It should be understood that the above are merely illustrative examples and do not constitute any limitation on the technical solutions of the present invention. In specific applications, those skilled in the art can make settings as needed, and the present invention does not impose any restrictions on this.

[0138] This embodiment detects abnormal signals in multi-source events, identifies abnormal events within those events based on these signals, acquires historical data of the abnormal events over a given time period, and determines a prediction model based on this data. The abnormal event is then input into the prediction model to determine the prediction result. Based on the prediction result, it is determined whether an early warning is needed. If an early warning is required, the abnormal event and the prediction result are input into a hierarchical model to determine the warning level, and an early warning is issued based on this level. This method enables real-time monitoring of abnormal signals in multi-source events, identification of abnormal events based on these signals, and prediction of these events using a prediction model to determine the need for an early warning. This allows for automatic analysis and early warning of multi-source data, improving analysis and warning efficiency while proactively identifying event risks, thus effectively preventing significant losses to the company due to abnormal events.

[0139] In one embodiment, the determining module 10 is further configured to:

[0140] Obtain high-frequency and low-frequency data for each event in the multi-source event;

[0141] By correlating the high-frequency data with the monitoring frequency of the low-frequency data, key information can be obtained;

[0142] An initial monitoring model is constructed based on the high-frequency data;

[0143] The initial monitoring model is estimated using the key information to obtain the monitoring model;

[0144] The multi-source events are input into the monitoring model to obtain monitoring results, wherein the monitoring results are used to determine whether abnormal signals occur in the multi-source events.

[0145] In one embodiment, the acquisition module 20 is further configured to:

[0146] Obtain historical data of the aforementioned abnormal event within the historical time period;

[0147] Based on the historical data, determine the duration of the abnormal event within the historical time period, and filter out the target duration period that is greater than the target threshold from the duration period.

[0148] Determine the feature vector of the abnormal event based on the target duration segment;

[0149] The feature vector is input into the initial prediction model to obtain the prediction model.

[0150] In one embodiment, the determining module 10 is further configured to:

[0151] Based on the historical data, determine the first probability that the abnormal event will trigger an early warning in the current time period;

[0152] Based on the historical data, a second probability of the abnormal event triggering an early warning within different periods is determined;

[0153] Establish a relationship curve between the historical data and the average number of warnings for the abnormal event within a historical time period, and determine a third possibility based on the relationship curve;

[0154] Based on the first possibility, the second possibility, and the third possibility, an initial prediction model is trained to obtain the hierarchical model;

[0155] The abnormal event and the prediction result are input into the hierarchical model to determine the warning level of the abnormal event.

[0156] In one embodiment, the warning levels include a first warning level, a second warning level, and a third warning level; wherein,

[0157] The early warning module 40 is also used for:

[0158] If the warning level is the first warning level, the warning of the abnormal event is completed by playing a voice message.

[0159] If the warning level is the second warning level, the warning of the abnormal event is completed through visual display;

[0160] If the warning level is the third warning level, the abnormal event is interpreted to obtain the interpretation result, and the interpretation result is displayed visually to complete the warning of the abnormal event.

[0161] In one embodiment, the step of interpreting the abnormal event to obtain an interpretation result and then visually displaying the interpretation result to complete the early warning of the abnormal event includes:

[0162] Statistical analysis and machine learning are used to identify patterns, trends, and outliers in the abnormal events, and the patterns, trends, and outliers are used as the interpretation results of the abnormal events.

[0163] The visualization format is determined, and the interpretation results are displayed visually based on the visualization format to complete the early warning of the abnormal event. The visualization format includes charts, graphs, and images.

[0164] In one embodiment, the early warning module 40 is further configured to:

[0165] After issuing an early warning for the abnormal event based on the warning level, the number of warnings for the abnormal event within a future time period is obtained.

[0166] A threshold number is set, and when the number of warnings exceeds the threshold number, a rapid warning mode for the abnormal event is activated;

[0167] When an abnormal event is detected as having an abnormal signal in the multi-source event while the abnormal event is in the rapid early warning mode, an early warning is issued.

[0168] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of this invention. In practical applications, those skilled in the art can select some or all of the workflow to achieve the purpose of this embodiment according to actual needs, and no restrictions are imposed here.

[0169] In addition, for technical details not described in detail in this embodiment, please refer to the event early warning method based on big data analysis provided in any embodiment of the present invention, which will not be repeated here.

[0170] Furthermore, it should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0171] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0172] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory (ROM) / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0173] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. An event early warning method based on big data analysis, characterized in that, The event early warning method based on big data analysis includes: When an abnormal signal is detected in a multi-source event, the abnormal event in the multi-source event is determined based on the abnormal signal; Obtain historical data of the abnormal event within a historical time period, and determine a prediction model based on the historical data; The abnormal event is input into the prediction model to determine the prediction result of the abnormal event; Determine whether an early warning is needed for the abnormal event based on the prediction results; When it is determined that an early warning for the abnormal event is required, the abnormal event and the prediction result are input into a hierarchical model to determine the early warning level of the abnormal event, and the early warning for the abnormal event is completed according to the early warning level. Specifically, based on historical data, a first probability of an early warning occurring in the current time period is determined; based on historical data, a second probability of an early warning occurring in different periods is determined; a relationship curve is established between the historical data and the average number of early warnings for the abnormal event over a historical time period, and a third probability is determined based on the relationship curve; based on the first probability, the second probability, and the third probability, an initial prediction model is trained to obtain the predicted outcome. A hierarchical model is used to input the abnormal event and the prediction result into the hierarchical model to determine the warning level of the abnormal event. The warning level includes a first warning level, a second warning level, and a third warning level. The warning process based on the warning level includes: if the warning level is the first warning level, issuing a warning for the abnormal event via voice playback; if the warning level is the second warning level, issuing a warning for the abnormal event via visual display; if the warning level is the third warning level, interpreting the abnormal event to obtain an interpretation result, and issuing a warning for the abnormal event via visual display of the interpretation result. Before determining the abnormal event among the multi-source events based on the abnormal signal when an abnormal signal is detected in a multi-source event, the method further includes: The process involves acquiring high-frequency and low-frequency data for each event in the multi-source event; correlating the high-frequency data with the monitoring frequency of the low-frequency data to obtain key information; constructing an initial monitoring model based on the high-frequency data; estimating the initial monitoring model using the key information to obtain a monitoring model; and inputting the multi-source event into the monitoring model to obtain monitoring results, wherein the monitoring results are used to determine whether any abnormal signals occur in the multi-source event.

2. The method as described in claim 1, characterized in that, The step of acquiring historical data of the abnormal event within a historical time period and determining a prediction model based on the historical data includes: Obtain historical data of the aforementioned abnormal event within the historical time period; Based on the historical data, determine the duration of the abnormal event within the historical time period, and filter out the target duration period that is greater than the target threshold from the duration period. Determine the feature vector of the abnormal event based on the target duration segment; The feature vector is input into the initial prediction model to obtain the prediction model.

3. The method as described in claim 1, characterized in that, The process of interpreting the abnormal event to obtain an interpretation result, and then visually displaying the interpretation result to provide an early warning for the abnormal event, includes: Statistical analysis and machine learning are used to identify patterns, trends, and outliers in the abnormal events, and the patterns, trends, and outliers are used as the interpretation results of the abnormal events. The visualization format is determined, and the interpretation results are displayed visually based on the visualization format to complete the early warning of the abnormal event. The visualization format includes charts, graphs, and images.

4. The method as described in claim 3, characterized in that, The event early warning method based on big data analysis also includes: After issuing an early warning for the abnormal event based on the warning level, the number of warnings for the abnormal event within a future time period is obtained. A threshold number is set, and when the number of warnings exceeds the threshold number, a rapid warning mode for the abnormal event is activated; When an abnormal event is detected as having an abnormal signal in the multi-source event while the abnormal event is in the rapid early warning mode, an early warning is issued.

5. An event early warning device based on big data analysis, characterized in that, The event early warning device based on big data analysis includes: The determination module is used to determine the abnormal event among the multi-source events based on the abnormal signal when an abnormal signal is detected in the multi-source events. The acquisition module is used to acquire historical data of the abnormal event within a historical time period and determine a prediction model based on the historical data. The determining module is used to input the abnormal event into the prediction model and determine the prediction result of the abnormal event; The judgment module is used to determine whether an early warning is needed for the abnormal event based on the prediction result. The early warning module is used to, when it is determined that an early warning for an abnormal event is needed, input the abnormal event and the prediction result into a hierarchical model to determine the early warning level of the abnormal event, and complete the early warning for the abnormal event based on the early warning level; determine a first probability that the abnormal event will trigger an early warning in the current time period based on historical data; determine a second probability that the abnormal event will trigger an early warning in different periods based on historical data; establish a relationship curve between the historical data and the average number of early warnings for the abnormal event in a historical time period, and determine a third probability based on the relationship curve; and train an initial prediction model based on the first probability, the second probability, and the third probability to obtain... The hierarchical model takes the abnormal event and the prediction result as inputs and determines the warning level of the abnormal event. The warning levels include a first warning level, a second warning level, and a third warning level. The warning process based on the warning level includes: if the warning level is the first warning level, issuing a warning for the abnormal event via voice playback; if the warning level is the second warning level, issuing a warning for the abnormal event via visual display; if the warning level is the third warning level, interpreting the abnormal event to obtain an interpretation result, and issuing a warning for the abnormal event via visual display of the interpretation result. The determining module is further configured to acquire high-frequency and low-frequency data of each event in the multi-source event; correlate the high-frequency data with the monitoring frequency of the low-frequency data to obtain key information; construct an initial monitoring model based on the high-frequency data; estimate the initial monitoring model using the key information to obtain a monitoring model; input the multi-source event into the monitoring model to obtain monitoring results, wherein the monitoring results are used to determine whether abnormal signals occur in the multi-source event.

6. An event early warning device based on big data analysis, characterized in that, The device includes: a memory, a processor, and a big data analytics-based event warning program stored in the memory and executable on the processor, the big data analytics-based event warning program being configured to implement the steps of the big data analytics-based event warning method as described in any one of claims 1 to 4.

7. A storage medium, characterized in that, The storage medium stores an event warning program based on big data analysis, which, when executed by a processor, implements the steps of the event warning method based on big data analysis as described in any one of claims 1 to 4.

Images