Automatic public opinion risk early warning application method

Abstract

The invention discloses an automatic public opinion risk early warning application method, and relates to the technical field of public opinion risk early warning management. The automatic public opinion risk early warning application method comprises the following steps: S1, early warning real-time evaluation; s2, data delay evaluation; and S3, performing data optimization evaluation. The early warning real-time evaluation value is obtained through the obtained public opinion data, whether data delay evaluation is carried out or not is judged based on the early warning real-time evaluation value, then if data delay evaluation is carried out, the data delay evaluation value is obtained according to the obtained initial early warning data, and whether early warning optimization is carried out or not is judged based on the data delay evaluation value. And finally, if early warning optimization is carried out, a data optimization evaluation value is obtained according to the initial early warning data and the early warning data after early warning optimization, and whether feedback is carried out is judged based on the data optimization evaluation value, so that the effect of improving the public opinion risk early warning efficiency is achieved. The problem of low early warning efficiency in a public opinion risk early warning application process in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of public opinion risk early warning management, and in particular to an automated public opinion risk early warning application method. Background Art

[0002] With the rapid development of network technology, the emergence of online self-media has accelerated the spread of public opinion information. In today's increasingly complex information dissemination environment, public opinion risk has become one of the important challenges faced by enterprises and governments. Once public opinion risk breaks out, it may cause serious damage to the brand image of the enterprise, the credibility of the government, and the reputation of individuals. Traditional public opinion risk warning methods often use fixed thresholds, that is, when a certain indicator (such as the number of visits, the number of comments, etc.) exceeds the preset threshold, the warning is triggered. Through the automated public opinion risk warning application method, changes in public opinion can be monitored and analyzed in real time, potential public opinion risks can be discovered in a timely manner, and corresponding countermeasures can be taken.

[0003] Existing technologies monitor and analyze various public opinion data sources in real time, extract potential risk signals from them, and evaluate and respond to public opinion risks, thereby achieving monitoring and early warning of public opinion risks.

[0004] For example, the public opinion analysis and financial early warning method, device, electronic device and storage medium announced by the invention patent with announcement number: CN111784492B include: collecting initial early warning data of the object to be analyzed; filtering out non-natural public opinion data in the initial early warning data based on the public opinion influence parameters, public opinion quality parameters and public opinion targeted parameters of each user in the initial early warning data to obtain natural public opinion data; and performing public opinion analysis on the object based on the natural public opinion data.

[0005] For example, the public opinion data processing method, device and storage medium announced in the invention patent with announcement number: CN112711705B include: obtaining target-related information from the public opinion data to be processed according to a preset knowledge graph; obtaining the sentiment tendency information and topic evolution information of the public opinion data to be processed according to the target-related information, and obtaining historical public opinion data similar to the public opinion data to be processed from a preset historical training set; determining the rating parameters of the public opinion data to be processed according to the sentiment tendency information, topic evolution information and historical public opinion data; obtaining the evaluation results of the public opinion data to be processed according to the rating parameters and a first preset threshold, and displaying the evaluation results in the corresponding interface according to the evaluation results, and displaying at least one of the sentiment tendency information, topic evolution information and historical public opinion data.

[0006] However, in the process of implementing the technical solution of the invention in the embodiments of the present application, the present application found that the above technology has at least the following technical problems:

[0007] In the prior art, due to the low real-time performance of information retrieval, early warning applications may not be able to obtain the latest public opinion information in a timely manner, which may cause the issuance of early warning signals to lag behind the occurrence of actual events, resulting in the problem of low early warning efficiency in the application of public opinion risk early warning. Summary of the invention

[0008] The embodiment of the present application solves the problem of low warning efficiency in the application process of public opinion risk warning in the prior art by providing an automated public opinion risk warning application method, thereby improving the efficiency of public opinion risk warning.

[0009] The embodiment of the present application provides an automated public opinion risk warning application method, comprising the following steps: S1, obtaining a real-time warning evaluation value based on the acquired public opinion data, and judging whether to perform a data delay evaluation based on the real-time warning evaluation value, wherein the real-time warning evaluation value is used to evaluate the real-time nature of the public opinion warning; S2, if a data delay evaluation is performed, obtaining a data delay evaluation value based on the acquired initial warning data, and judging whether to perform warning optimization based on the data delay evaluation value, wherein the data delay evaluation value is used to evaluate the delay in data collection and processing; S3, if warning optimization is performed, obtaining a data optimization evaluation value based on the initial warning data and the warning data after warning optimization, and judging whether to perform feedback based on the data optimization evaluation value, wherein the data optimization evaluation value is used to evaluate the effectiveness of the optimization measures.

[0010] Furthermore, the public opinion data includes keyword discussion volume, warning keyword discussion volume, keyword appearance time and warning time; the initial warning data includes Ping value, packet loss rate, bandwidth, content update frequency, CPU utilization, memory utilization, I / O waiting time and throughput.

[0011] Furthermore, the specific method for obtaining the real-time evaluation value of the early warning based on the acquired public opinion data is as follows: obtaining a preset early warning interval duration from a preset database; obtaining a discussion volume deviation based on the relative relationship between the keyword discussion volume and the early warning keyword discussion volume, the discussion volume deviation being obtained by processing the ratio operation of the keyword discussion volume and the early warning keyword discussion volume; obtaining the early warning delay based on the relative relationship between the keyword appearance time, the early warning time and the preset early warning interval duration, the early warning delay being obtained by processing the ratio operation of the keyword appearance time and the early warning time; processing the discussion volume deviation and the early warning delay to obtain the real-time evaluation value of the early warning.

[0012] Furthermore, the specific method for obtaining the data delay evaluation value according to the obtained initial warning data is as follows: obtaining preset warning data from a preset database, the preset warning data including a preset Ping value, a preset data packet loss rate, a preset I / O waiting time, a preset CPU utilization maximum value, and a preset memory utilization maximum value; obtaining a Ping value deviation according to a relative relationship between the Ping value and the preset Ping value, the Ping value deviation being obtained by processing a ratio operation of the Ping value and the preset Ping value; obtaining a data packet loss rate deviation according to a relative relationship between the data packet loss rate and the preset data packet loss rate, the data packet loss rate deviation being obtained by processing a ratio operation of the data packet loss rate and the preset data packet loss rate; obtaining an update coefficient according to a relative relationship between bandwidth and content update frequency, the update coefficient being obtained by processing a ratio operation of the bandwidth and content update frequency. Processing is performed to obtain; a CPU utilization deviation is obtained according to a relative relationship between the CPU utilization and a preset maximum CPU utilization, and the CPU utilization deviation is obtained by processing a ratio operation between the CPU utilization and the preset maximum CPU utilization; a memory utilization deviation is obtained according to a relative relationship between the memory utilization and the preset maximum memory utilization, and the memory utilization deviation is obtained by processing a ratio operation between the memory utilization and the preset maximum memory utilization; an I / O waiting time deviation is obtained according to a relative relationship between the I / O waiting time and the preset I / O waiting time, and the I / O waiting time deviation is obtained by processing a ratio operation between the I / O waiting time and the preset I / O waiting time; the Ping value deviation, the packet loss rate deviation, the update coefficient, the CPU utilization deviation, the memory utilization deviation and the I / O waiting time deviation are processed to obtain a data delay evaluation value.

[0013] Furthermore, the specific method for obtaining the data optimization evaluation value according to the initial warning data and the warning data after warning optimization is as follows: obtaining a Ping value optimization coefficient according to the relative relationship between the optimized Ping value and the Ping value, and the Ping value optimization coefficient is obtained by processing the ratio operation of the optimized Ping value and the Ping value; obtaining a packet loss rate optimization coefficient according to the relative relationship between the optimized packet loss rate and the packet loss rate, and the packet loss rate optimization coefficient is obtained by processing the ratio operation of the optimized packet loss rate and the packet loss rate; obtaining a bandwidth optimization coefficient according to the relative relationship between the optimized bandwidth and the bandwidth, and the bandwidth optimization coefficient is obtained by processing the ratio operation of the optimized bandwidth and the bandwidth; obtaining a throughput optimization coefficient according to the relative relationship between the optimized throughput, the throughput and the preset throughput, and the throughput optimization coefficient is obtained by processing the ratio operation of the optimized throughput to the preset throughput and the ratio operation of the throughput to the preset throughput; Pi The ng value optimization coefficient, the packet loss rate optimization coefficient, the bandwidth optimization coefficient and the throughput optimization coefficient are processed to obtain the network optimization effective coefficient, and the network optimization effective coefficient is used to evaluate the effectiveness of network optimization; the CPU utilization optimization coefficient is obtained according to the relative relationship between the optimized CPU utilization deviation and the CPU utilization deviation, and the CPU utilization optimization coefficient is obtained by processing the difference between the optimized CPU utilization deviation and the CPU utilization deviation; the memory utilization optimization coefficient is obtained according to the relative relationship between the optimized memory utilization deviation and the memory utilization deviation, and the memory utilization optimization coefficient is obtained by processing the difference between the optimized memory utilization deviation and the memory utilization deviation; the I / O waiting time optimization coefficient is obtained according to the relative relationship between the optimized I / O waiting time and the I / O waiting time, and the I / O waiting time optimization coefficient is obtained by processing the ratio of the optimized I / O waiting time to the I / O waiting time; the CPU utilization optimization coefficient, the memory utilization optimization coefficient and the I / O waiting time optimization coefficient are processed to obtain the resource optimization effective coefficient; the network optimization effective coefficient and the resource optimization effective coefficient are processed to obtain the data optimization evaluation value.

[0014] Furthermore, the specific limiting expression of the data optimization evaluation value is as follows:

[0015]

[0016] In the formula, i represents the number of times of public opinion monitoring, i=1,2,...,T, T represents the total number of times of public opinion monitoring, YOU i represents the data optimization evaluation value of the ith public opinion monitoring, WL i represents the network optimization effectiveness coefficient of the ith public opinion monitoring, ZY i It represents the resource optimization efficiency coefficient of the i-th public opinion monitoring.

[0017] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0018] 1. The real-time evaluation value of the early warning is obtained by the acquired public opinion data, and then the data delay evaluation value obtained based on the initial early warning data is used to determine whether to optimize the early warning. Finally, if data optimization evaluation is performed, the data optimization evaluation value is obtained based on the initial early warning data and the early warning data after early warning optimization. Based on the data optimization evaluation value, it is determined whether to provide feedback, thereby improving the real-time nature of the early warning, and further realizing the improvement of the efficiency of public opinion risk early warning, which effectively solves the problem of low early warning efficiency in the application process of public opinion risk early warning in the prior art.

[0019] 2. The discussion volume deviation is obtained through the relative relationship between the keyword discussion volume and the warning keyword discussion volume. Then, the warning delay is obtained based on the relative relationship between the keyword appearance time, the warning time and the preset warning interval. Finally, the discussion volume deviation and the warning delay are processed to obtain the real-time evaluation value of the warning, thereby quantitatively evaluating the real-time nature of the public opinion risk warning, and then realizing the optimization of the public opinion risk warning mechanism during the application process.

[0020] 3. The network optimization effectiveness coefficient is obtained by processing the Ping value optimization coefficient, the packet loss rate optimization coefficient, the bandwidth optimization coefficient and the throughput optimization coefficient. Then, the CPU utilization optimization coefficient, the memory utilization optimization coefficient and the I / O waiting time optimization coefficient are processed to obtain the resource optimization effectiveness coefficient. Finally, the network optimization effectiveness coefficient and the resource optimization effectiveness coefficient are processed to obtain the data optimization evaluation value, thereby quantitatively evaluating the effectiveness of the optimization measures and improving the real-time performance of public opinion risk warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A flowchart of an automated public opinion risk warning application method provided in an embodiment of the present application;

[0022] Figure 2 A schematic diagram showing how the network optimization effectiveness coefficient provided in the embodiment of the present application changes with the Ping value after optimization;

[0023] Figure 3 A schematic diagram showing the change of the network optimization effectiveness coefficient provided in the embodiment of the present application as the optimized data packet loss rate changes;

[0024] Figure 4 A schematic diagram showing the change of the network optimization effectiveness coefficient according to the optimized bandwidth provided in the embodiment of the present application;

[0025] Figure 5A schematic diagram of the change of the network optimization effectiveness coefficient provided in an embodiment of the present application as the throughput after optimization changes. DETAILED DESCRIPTION

[0026] The embodiment of the present application solves the problem of low warning efficiency in the application process of public opinion risk warning in the prior art by providing an automated public opinion risk warning application method. A real-time warning evaluation value is obtained through the acquired public opinion data, and whether to perform data delay evaluation is determined based on the real-time warning evaluation value. Then, a data delay evaluation value is obtained based on the acquired initial warning data, and whether to perform warning optimization is determined based on the data delay evaluation value. Finally, a data optimization evaluation value is obtained based on the initial warning data and the warning data after warning optimization, and whether to provide feedback is determined based on the data optimization evaluation value, thereby improving the efficiency of public opinion risk warning.

[0027] The technical solution in the embodiment of the present application is to solve the problem of low warning efficiency in the application process of the above-mentioned public opinion risk warning. The overall idea is as follows:

[0028] The real-time evaluation value of the warning is obtained by the acquired public opinion data, and then the data delay evaluation value obtained based on the initial warning data is used to determine whether to perform warning optimization. Finally, if data optimization evaluation is performed, the data optimization evaluation value is obtained based on the initial warning data and the warning data after warning optimization. Based on the data optimization evaluation value, it is determined whether to provide feedback, thereby achieving the effect of improving the efficiency of public opinion risk warning.

[0029] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0030] like Figure 1 As shown, it is a flowchart of an automated public opinion risk warning application method provided in an embodiment of the present application, and the method includes the following steps: S1, warning real-time evaluation: obtain a warning real-time evaluation value based on the acquired public opinion data, and judge whether to perform data delay evaluation based on the warning real-time evaluation value, and the warning real-time evaluation value is used to evaluate the real-time nature of the public opinion warning; S2, data delay evaluation: if data delay evaluation is performed, obtain a data delay evaluation value based on the acquired initial warning data, and judge whether to perform warning optimization based on the data delay evaluation value, and the data delay evaluation value is used to evaluate the delay in data collection and processing; S3, data optimization evaluation: if warning optimization is performed, obtain a data optimization evaluation value based on the initial warning data and the warning data after warning optimization, and judge whether to perform feedback based on the data optimization evaluation value, and the data optimization evaluation value is used to evaluate the effectiveness of the optimization measures.

[0031] It should be added that the public opinion data includes the amount of keyword discussion, the amount of warning keyword discussion, the time when the keyword appears and the warning time; the initial warning data includes Ping value, packet loss rate, bandwidth, content update frequency, CPU utilization, memory utilization, I / O waiting time and throughput.

[0032] It should be added that it is determined whether the early warning real-time evaluation value is less than the preset early warning real-time threshold. If the early warning real-time evaluation value is less than the preset early warning real-time threshold, data delay evaluation is performed, otherwise data delay evaluation is not performed.

[0033] It should be added that it is determined whether the data optimization evaluation value is less than the preset optimization threshold. If the data optimization evaluation value is less than the preset optimization threshold, feedback is given to the preset personnel, otherwise no feedback is given to the preset personnel.

[0034] In this embodiment, the keyword discussion volume is obtained by acquiring content containing keywords in real time through the social media API and performing statistics; the warning keyword discussion volume represents the keyword discussion volume before the warning moment; the keyword appearance moment records the time when the keyword first appears through a timestamp; the warning moment is recorded through a timestamp.

[0035] The Ping command is used to obtain the Ping value, the total number of data packets, and the number of data packet losses; the data packet loss rate is obtained by performing a ratio operation on the number of data packet losses and the total number of data packets; the bandwidth of the current network is obtained by using bandwidth monitoring tools (such as Iperf, Ntop, and Wireshark); the content update frequency is obtained by checking the website's update log or using a website content crawler; the CPU utilization is obtained by operating application monitoring commands (such as Linux's top command and htop command) or monitoring tools (such as Prometheus and Zabbix); the memory utilization is obtained by operating application monitoring commands (such as Linux's free command and vmstat command) or monitoring tools; the I / O waiting time is obtained by operating application monitoring commands (such as Linux's iostat command) or dedicated monitoring tools (such as Nagios and Zabbix); the throughput is obtained by obtaining the number of data packets or data volume transmitted per second through monitoring tools (such as Iperf and Wireshark) and performing statistics.

[0036] The preset warning real-time threshold is represented by the minimum value of the warning real-time evaluation value in the historical time period; the preset optimization threshold is represented by the maximum value of the data optimization evaluation value in the historical time period.

[0037] Through the above steps, the efficiency of public opinion risk warning is improved, and the problem of low warning efficiency in the application of public opinion risk warning in the existing technology is effectively solved; by obtaining the discussion volume of keywords and warning keywords in real time, the public's attention to a certain event or topic can be timely understood, providing data support for public opinion risk warning; by monitoring the initial warning data, the network performance and operation status can be comprehensively evaluated, so as to timely discover and solve problems.

[0038] Furthermore, the specific method for obtaining the early warning real-time evaluation value based on the obtained public opinion data is as follows: obtain the preset early warning interval from the preset database; obtain the discussion volume deviation based on the relative relationship between the keyword discussion volume and the early warning keyword discussion volume (that is, the specific restriction expression of the early warning real-time evaluation value is ), the discussion volume deviation is obtained by calculating the ratio of the keyword discussion volume to the warning keyword discussion volume; the warning delay (i.e., the specific restriction expression of the warning real-time evaluation value) is obtained according to the relative relationship between the keyword appearance time, the warning time, and the preset warning interval. ), the warning delay is obtained by processing the ratio of the keyword appearance time and the warning time; the discussion volume deviation and the warning delay are processed to obtain the warning real-time evaluation value.

[0039] The specific limiting expression of the early warning real-time evaluation value is as follows:

[0040]

[0041] In the formula, i represents the number of public opinion monitoring times, i = 1, 2, ..., T, T represents the total number of public opinion monitoring times, p represents the number of warning times, p = 1, 2, ..., P, P represents the total number of warning times, represents the discussion volume of the keyword in the pth warning in the ith public opinion monitoring, represents the discussion volume of the warning keyword of the pth warning in the ith public opinion monitoring, Indicates the time when the keyword of the pth warning appears in the i-th public opinion monitoring, represents the warning time of the pth warning in the ith public opinion monitoring, SKE0 represents the preset warning interval, and α represents the discussion volume influencing factor.

[0042] In this embodiment, the preset warning interval is set according to the risk rating of the keyword. For example, for the keyword "parking" with a low risk rating, it can be set to 1440 minutes (i.e., 24 hours), and for the keyword "unreasonable charges" with a high risk rating, it can be set to 60 minutes; the warning keyword discussion volume represents the keyword discussion volume at the warning time.

[0043] The greater the discussion volume deviation, the greater the difference between the keyword discussion volume and the warning keyword discussion volume, which may make it difficult for the warning application to accurately judge the actual degree and urgency of the risk, thereby increasing the risk of warning delay; the greater the discussion volume deviation, the smaller the real-time evaluation value of the warning may be; the increase in warning delay may lead to further development of the warning delay problem if it is not discovered and handled in time, thereby increasing the discussion volume deviation; the greater the warning delay, the smaller the real-time evaluation value of the warning may be, indicating that the real-time nature of the public opinion risk warning is worse.

[0044] definition Among them, TA i.p represents the discussion volume deviation of the p-th warning in the i-th public opinion monitoring, and defines Among them, SK i.p It represents the warning delay of the pth warning in the i-th public opinion monitoring. Taking the total number of public opinion monitoring as 1, the total number of warnings as 1, α as 1.145, and the preset warning interval as 60 minutes as an example, the change statistics of the real-time evaluation value of the warning are shown in Table 1:

[0045] Table 1 Statistics of changes in early warning real-time evaluation values

[0046]

[0047] It can be seen from the first and second groups of data in the table that when the discussion volume deviation increases, the real-time evaluation value of the warning decreases accordingly; it can be seen from the second and third groups of data that when the warning delay increases, the real-time evaluation value of the warning decreases accordingly; through the above steps, a quantitative evaluation of the real-time nature of public opinion risk warning is achieved, thereby optimizing the public opinion risk warning mechanism.

[0048] Further, a specific method for obtaining a data delay evaluation value according to the obtained initial warning data is as follows: obtaining preset warning data from a preset database, the preset warning data including a preset Ping value, a preset packet loss rate, a preset I / O waiting time, a preset CPU utilization maximum value, and a preset memory utilization maximum value; obtaining a Ping value deviation (i.e., the value of the specific restriction expression of the data delay evaluation value) according to the relative relationship between the Ping value and the preset Ping value; ), the Ping value deviation is obtained by processing the ratio of the Ping value to the preset Ping value; the packet loss rate deviation (i.e., the specific restriction expression of the data delay evaluation value) is obtained according to the relative relationship between the packet loss rate and the preset packet loss rate. ), the packet loss rate deviation is obtained by processing the ratio of the packet loss rate to the preset packet loss rate; the update coefficient (i.e., the specific restriction expression of the data delay evaluation value) is obtained according to the relative relationship between the bandwidth and the content update frequency. ), the update coefficient is obtained by processing the ratio of bandwidth and content update frequency; the CPU utilization deviation (i.e., the CPUP in the specific restriction expression of the data delay evaluation value) is obtained according to the relative relationship between the CPU utilization and the preset maximum CPU utilization. i ), the CPU utilization deviation is obtained by calculating the ratio of the CPU utilization to the preset maximum CPU utilization; the memory utilization deviation (i.e., the NEIP in the specific restriction expression of the data delay evaluation value) is obtained according to the relative relationship between the memory utilization and the preset maximum memory utilization. i ), the memory utilization deviation is obtained by calculating the ratio of the memory utilization to the preset maximum memory utilization; the I / O waiting time deviation (i.e., the specific restriction expression of the data delay evaluation value) is obtained according to the relative relationship between the I / O waiting time and the preset I / O waiting time. ), the I / O waiting time deviation is obtained by processing the ratio of the I / O waiting time and the preset I / O waiting time; the Ping value deviation, the packet loss rate deviation, the update coefficient, the CPU utilization deviation, the memory utilization deviation and the I / O waiting time deviation are processed to obtain the data delay evaluation value.

[0049] The specific limiting expression of the data delay evaluation value is as follows:

[0050]

[0051] Where i represents the number of times of public opinion monitoring, i=1,2,...,T, T represents the total number of times of public opinion monitoring, WLY i represents the data delay evaluation value of the ith public opinion monitoring, PING i represents the Ping value of the ith public opinion monitoring, PING0 represents the preset Ping value, DAIK i represents the bandwidth of the ith public opinion monitoring, NRGX i represents the content update frequency of the ith public opinion monitoring, DIUS i represents the packet loss rate of the ith public opinion monitoring, DIUS0 represents the preset packet loss rate, a is the traffic impact factor, CPUP i represents the CPU utilization deviation of the ith public opinion monitoring, NEIP i represents the memory utilization deviation of the ith public opinion monitoring, IOSJ i represents the I / O waiting time of the ith public opinion monitoring, IOSJ0 represents the preset I / O waiting time, and CPU i represents the CPU utilization of the ith public opinion monitoring, CPU max Indicates the preset maximum CPU utilization, NEI irepresents the memory utilization of the ith public opinion monitoring, NEI max Indicates the preset maximum memory usage.

[0052] In this embodiment, the preset Ping value is represented by the average value of the Ping value in the historical time period; the preset packet loss rate is represented by the average value of the packet loss rate in the historical time period; the preset I / O waiting time is represented by the average value of the I / O waiting time in the historical time period; the preset CPU utilization maximum value is set according to the application performance, for example, the standard CPU utilization range is 50%-80%, then the preset CPU utilization maximum value is set to 80%; the preset memory utilization maximum value is set according to the application performance, for example, the standard memory utilization range is 50%-80%, then the preset memory utilization maximum value is set to 80%; through the above steps, a quantitative evaluation of data delay is achieved, thereby achieving improved operation stability.

[0053] A large deviation in the Ping value may lead to an increase in the deviation in the packet loss rate, because network instability will increase the risk of packet loss. When the Ping value deviation increases, the I / O waiting time may also increase, because the transmission time of the data packet in the network increases. Higher CPU utilization may cause the application to process network requests more slowly, thereby increasing the Ping value. The greater the deviation in the Ping value, the greater the data delay assessment value. The greater the deviation in the packet loss rate, the greater the data delay assessment value.

[0054] Higher CPU utilization may lead to increased memory utilization, because data is frequently read, written, and stored. As memory utilization increases, memory utilization deviation will also increase. Conversely, an increase in memory utilization deviation may also lead to an increase in CPU utilization deviation. When CPU utilization is higher, since applications need to frequently access resources of hard disks, networks, or other external devices, it may lead to: a longer I / O wait time; a larger CPU utilization deviation leads to a larger data delay assessment value; a larger memory utilization deviation leads to a larger data delay assessment value; a larger I / O wait time leads to a larger data delay assessment value.

[0055] An increase in the frequency of content updates may mean an increase in user activities in social networks, which may lead to an increase in application load, and in turn affect the increase in CPU utilization, memory utilization, etc.; the larger the bandwidth, the stronger the network's ability to transmit data in theory, and the smaller the possibility of data packet loss and delay; the greater the frequency of content updates, the greater the data delay assessment value; the larger the bandwidth, the smaller the data delay assessment value, which means that the delay in data collection and processing is smaller.

[0056] Furthermore, the specific process of determining whether to perform early warning optimization based on the data delay evaluation value is as follows: Determine whether the data delay evaluation value is less than the preset delay threshold: If the data delay evaluation value is less than the preset delay threshold, no early warning optimization is performed; If the data delay evaluation value is not less than the preset delay threshold, early warning optimization is performed; Early warning optimization includes network optimization and resource optimization.

[0057] In this embodiment, the preset delay threshold is the maximum value of the data delay evaluation value in the historical time period; through the above steps, the data transmission and processing process are dynamically adjusted to ensure the real-time and accuracy of the early warning. When the data delay is high, the application is automatically optimized to reduce the delay, ensuring that the early warning application responds to changes in public opinion risks in a timely manner and improving the adaptability of the early warning application under high delay conditions.

[0058] Furthermore, the specific process of network optimization is as follows: network optimization includes bandwidth optimization, content distribution network acceleration, multi-threaded crawling and load balancing; bandwidth optimization includes increasing network bandwidth and data compression; content distribution network acceleration includes setting cache strategies for static content and cache strategies for dynamic content.

[0059] In this embodiment, the network bandwidth is increased by purchasing or upgrading network equipment to increase the physical bandwidth of the network, for example, upgrading the network connection from 100Mbps to 1Gbps; data compression reduces the size of transmitted data by using data compression technology (such as HTTP / 2 header compression, Gzip, etc.); multi-threaded crawling means dividing the crawler task into multiple threads, and each thread executes the crawling task independently; load balancing means sending the request to the server with the least number of current connections through the least connection algorithm to balance the load.

[0060] Static content usually includes pictures, videos, etc.; set an appropriate cache time according to the update frequency of the static content. For example, for pictures, you can set a cache time of 30 days; common cache strategies include full cache, conditional cache, and slice cache. For static content, the full cache strategy is usually used.

[0061] Dynamic content usually includes user comments, real-time data, etc.; for example, for user comments, a cache time of 10 minutes or 2 hours can be set to ensure the real-time nature of the data; for dynamic content with high real-time requirements, edge computing technology is used to process real-time data and cache it on content delivery network (CDN) nodes; edge computing technology pushes computing and data storage to the edge of the network, thereby reducing dependence on central servers and reducing network latency.

[0062] Through bandwidth optimization, the efficiency and speed of data transmission are improved, and transmission delays and bandwidth costs are reduced; by deploying nodes and caching content globally, user access speed is improved, and delays and bandwidth consumption are reduced; through multi-threaded crawling, the work efficiency of crawlers is improved, the crawling time is shortened, and the speed and accuracy of data acquisition are improved; through load balancing, the normal operation of all servers is ensured and resources are fully utilized, thereby improving the stability and scalability of the application.

[0063] Furthermore, the specific process of resource optimization is as follows: resource optimization includes algorithm reconstruction, multi-threaded optimization, use of asynchronous I / O, reduction of CPU frequency and memory management optimization; memory management optimization includes memory allocation strategy and use of memory pool.

[0064] In this embodiment, the risk rating algorithm includes a deep learning model, a multi-task learning model and a dynamic risk index algorithm; the algorithm reconstruction is achieved by eliminating unused variables, then merging code segments that calculate the same result multiple times, and then using compound assignment operators (such as +=, -=, etc.), and finally using reference passing or pointer passing to improve the efficiency of the risk rating algorithm; multi-threaded optimization means using multi-threading technology to decompose a task into multiple subtasks for parallel execution to improve the execution efficiency and response speed of the program.

[0065] In Linux systems, you can use the POSIX aio_* function series or the epoll mechanism to implement asynchronous I / O; in Windows systems, you can use IOCP (I / O Completion Port) to implement asynchronous I / O; you can also use third-party libraries (such as Boost.Asio) to simplify the implementation of asynchronous I / O.

[0066] Dynamically adjust the CPU frequency according to the needs of the application. For example, increase the CPU frequency when the keyword discussion volume is greater than the preset keyword discussion volume, otherwise reduce the CPU frequency; the preset keyword discussion volume is represented by the maximum value of the keyword discussion volume in the historical time period.

[0067] Memory allocation strategy means choosing an appropriate memory allocation strategy based on actual needs. For example, for small blocks of memory that need to be frequently allocated and released, use memory pools or object pools to reduce the overhead of memory allocation and release. For large data structures, use dynamic memory allocation to allocate and release memory as needed. Memory pool refers to a technology that pre-allocates a large memory area as a memory pool. When memory needs to be allocated, a small piece of memory is allocated from the pool. When it is no longer needed, the memory is returned to the pool instead of being released to the operating system. This reduces the frequency of memory allocation and release and improves memory usage efficiency.

[0068] Through the above steps, the stability of the public opinion risk warning application has been improved.

[0069] Further, the specific method of obtaining the data optimization evaluation value based on the initial warning data and the warning data after warning optimization is as follows: according to the relative relationship between the optimized Ping value and the Ping value, the Ping value optimization coefficient (that is, the specific restriction expression of the data optimization evaluation value) is obtained. ), the Ping value optimization coefficient is obtained by processing the ratio of the optimized Ping value and the Ping value; the packet loss rate optimization coefficient (i.e., the specific restriction expression of the data optimization evaluation value) is obtained according to the relative relationship between the optimized packet loss rate and the packet loss rate. ), the packet loss rate optimization coefficient is obtained by processing the ratio of the optimized packet loss rate and the packet loss rate; the bandwidth optimization coefficient (i.e., the specific restriction expression of the data optimization evaluation value) is obtained according to the relative relationship between the optimized bandwidth and the bandwidth ), the bandwidth optimization coefficient is obtained by processing the ratio of the optimized bandwidth to the bandwidth; the throughput optimization coefficient (i.e., the specific restriction expression of the data optimization evaluation value) is obtained according to the relative relationship between the optimized throughput, the throughput and the preset throughput. The throughput optimization coefficient is obtained by processing the ratio of the optimized throughput to the preset throughput and the ratio of the throughput to the preset throughput; the Ping value optimization coefficient, the packet loss rate optimization coefficient, the bandwidth optimization coefficient and the throughput optimization coefficient are processed to obtain the network optimization effective coefficient (that is, the specific restriction expression of the data optimization evaluation value WL i ), the network optimization effectiveness coefficient is used to evaluate the effectiveness of network optimization; the CPU utilization optimization coefficient (that is, the CPUP in the specific restriction expression of the data optimization evaluation value) is obtained according to the relative relationship between the optimized CPU utilization deviation and the CPU utilization deviation. i -CPUP i ′), the CPU utilization optimization coefficient is obtained by processing the difference between the optimized CPU utilization deviation and the CPU utilization deviation; the memory utilization optimization coefficient (i.e., the NEIP in the specific restriction expression of the data optimization evaluation value) is obtained according to the relative relationship between the optimized memory utilization deviation and the memory utilization deviation. i -NEIP i ′), the memory utilization optimization coefficient is obtained by processing the difference between the optimized memory utilization deviation and the memory utilization deviation; the I / O waiting time optimization coefficient (i.e., the specific restriction expression of the data optimization evaluation value) is obtained according to the relative relationship between the optimized I / O waiting time and the I / O waiting time. ), the I / O waiting time optimization coefficient is obtained by processing the ratio of the optimized I / O waiting time and the I / O waiting time; the CPU utilization optimization coefficient, the memory utilization optimization coefficient and the I / O waiting time optimization coefficient are processed to obtain the resource optimization effective coefficient (that is, the specific restriction expression of the data optimization evaluation value ZY i ) ; the network optimization effectiveness coefficient and the resource optimization effectiveness coefficient are processed to obtain the data optimization evaluation value.

[0070] Among them, the specific restriction expression of the data optimization evaluation value is as follows:

[0071]

[0072]

[0073] In the formula, i represents the number of times of public opinion monitoring, i=1,2,...,T, T represents the total number of times of public opinion monitoring, YOU i represents the data optimization evaluation value of the ith public opinion monitoring, WL i represents the network optimization effectiveness coefficient of the ith public opinion monitoring, ZY i represents the resource optimization efficiency coefficient of the i-th public opinion monitoring, PING i ′ represents the optimized Ping value of the ith public opinion monitoring, PING i represents the Ping value of the ith public opinion monitoring, DIUS i ′ represents the optimized packet loss rate of the ith public opinion monitoring, DIUS i represents the packet loss rate of the ith public opinion monitoring, DAIK i ′ represents the optimized bandwidth of the ith public opinion monitoring, DAIK i represents the bandwidth of the ith public opinion monitoring, TUNL i represents the throughput of the ith public opinion monitoring, TUNL0 represents the preset throughput, and TUNL i ′ represents the optimized throughput of the ith public opinion monitoring, CPUP i ′ represents the CPU utilization deviation after optimization of the i-th public opinion monitoring, CPUP i represents the CPU utilization deviation of the ith public opinion monitoring, NEIP i ′ represents the optimized memory utilization deviation of the ith public opinion monitoring, NEIP i represents the memory utilization deviation of the ith public opinion monitoring, IOSJ i ′ represents the optimized I / O waiting time of the ith public opinion monitoring, IOSJ i represents the I / O waiting time of the ith public opinion monitoring, CPU i′ represents the optimized CPU utilization rate of the ith public opinion monitoring, CPU max Indicates the preset maximum CPU utilization, NEI i ′ represents the optimized memory utilization of the ith public opinion monitoring, NEI max Indicates the preset maximum memory usage.

[0074] In this embodiment, an increase in bandwidth helps to reduce the Ping value, because a higher bandwidth means that data can be transmitted faster; an increase in throughput also helps to reduce the Ping value, because more data can be transmitted in a shorter time; higher bandwidth and throughput help to reduce the packet loss rate, because more data can be successfully transmitted in a shorter time; an increase in memory utilization also helps to improve CPU utilization and performance; optimizing I / O waiting time helps to improve CPU and memory utilization; the larger the Ping value optimization coefficient, the packet loss rate optimization coefficient, the throughput optimization coefficient and the I / O waiting time optimization coefficient, the smaller the data optimization evaluation value; the larger the bandwidth optimization coefficient, the CPU utilization optimization coefficient and the memory utilization optimization coefficient, the larger the data optimization evaluation value, which indicates that the effectiveness of the early warning optimization measures is higher.

[0075] For example, if the Ping value is 500 milliseconds, the packet loss rate is 0.5%, the bandwidth is 50Mbps, the throughput is 250Mbps, and the preset throughput is 500Mbps, Figure 2 , which is a schematic diagram of the change of the network optimization effectiveness coefficient provided in the embodiment of the present application with the change of the optimized Ping value. When the optimized packet loss rate is 0.1 Mbps, the optimized bandwidth is 60 Mbps, and the optimized throughput is 300 Mbps, the network optimization effectiveness coefficient decreases with the increase of the optimized Ping value;

[0076] like Figure 3 , which is a schematic diagram of the change of the network optimization effectiveness coefficient provided by the embodiment of the present application with the change of the optimized data packet loss rate. When the optimized Ping value is 10 milliseconds, the optimized bandwidth is 60Mbps, and the optimized throughput is 300Mbps, the network optimization effectiveness coefficient decreases with the increase of the optimized data packet loss rate;

[0077] like Figure 4 As shown, it is a schematic diagram of the change of the network optimization effectiveness coefficient provided by the embodiment of the present application with the change of the optimized bandwidth. When the optimized Ping value is 10 milliseconds, the optimized data packet loss rate is 0.1Mbps, and the optimized throughput is 300Mbps, the network optimization effectiveness coefficient increases with the increase of the optimized bandwidth;

[0078] like Figure 5As shown, it is a schematic diagram of the change of the network optimization effectiveness coefficient provided in the embodiment of the present application with the change of the optimized throughput. When the optimized Ping value is 10 milliseconds, the optimized data packet loss rate is 0.1Mbps, and the optimized bandwidth is 60Mbps, the network optimization effectiveness coefficient increases with the increase of the optimized throughput.

[0079] By comparing the data before and after optimization, we can achieve a quantitative assessment of the degree of improvement in network performance, resource utilization, etc., thereby improving the real-time nature of public opinion risk warnings.

[0080] Furthermore, it also includes conducting public opinion risk warnings, and the specific steps are: B1, setting the preset media website URL and noting the website name; B2, setting preset keywords and setting risk ratings; B3, searching according to the preset website URL and preset keywords, saving the URL corresponding to the preset keywords, and forming risk rating labels based on the preset keywords.

[0081] In this embodiment, the public opinion risk warning indicates predicting potential risk events or negative emotions by analyzing the public opinion data; the remarked website name indicates the name set for each website based on the preset media website URL; the preset media website URL is set according to business needs, such as "https: / / www.xiaohongshu.com", and the remark name is "Little Red Book"; the risk rating is set according to the importance of the keyword; the preset keywords are set according to business needs, such as "parking" is set to a low risk rating and "unreasonable charges" is set to a high risk rating; the retrieved URLs and content containing the preset keywords are saved to the local database; by implementing the public opinion risk warning steps, real-time monitoring and warning of sensitive information on specific media websites are realized, thereby improving the accuracy and efficiency of monitoring.

[0082] To summarize, the embodiment of the present application obtains a real-time warning evaluation value through the acquired public opinion data, and then determines whether to perform warning optimization based on the data delay evaluation value obtained from the acquired initial warning data. Finally, if warning optimization is performed, a data optimization evaluation value is obtained based on the initial warning data and the warning data after warning optimization, and whether to provide feedback is determined based on the data optimization evaluation value, thereby improving the real-time nature of the warning, and further achieving an improvement in the efficiency of public opinion risk warning, effectively solving the problem of low warning efficiency in the application process of public opinion risk warning in the prior art.

[0083] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0084] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0085] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0086] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0087] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0088] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. An automated public opinion risk early warning application method, characterized in that: The following steps are involved: S1, obtaining a warning real-time evaluation value according to the acquired public opinion data, and judging whether to perform data delay evaluation based on the warning real-time evaluation value, wherein the warning real-time evaluation value is used to evaluate the real-time nature of public opinion warning; S2, if data delay evaluation is performed, a data delay evaluation value is obtained according to the acquired initial warning data, and whether to perform warning optimization is determined based on the data delay evaluation value, wherein the data delay evaluation value is used to evaluate the delay in data collection and processing; S3: If early warning optimization is performed, a data optimization evaluation value is obtained based on the initial early warning data and the early warning data after early warning optimization, and whether to provide feedback is determined based on the data optimization evaluation value. The data optimization evaluation value is used to evaluate the effectiveness of the optimization measures.

2. The automated public opinion risk early warning application method as claimed in claim 1, characterized in that: The public opinion data includes the amount of keyword discussion, the amount of warning keyword discussion, the time when the keyword appeared and the warning time; The initial warning data includes Ping value, packet loss rate, bandwidth, content update frequency, CPU utilization, memory utilization, I / O waiting time and throughput.

3. The automated public opinion risk early warning application method as claimed in claim 2, characterized in that: The specific method of obtaining the real-time evaluation value of the early warning according to the acquired public opinion data is as follows: Obtain the preset warning interval duration from the preset database; The discussion volume deviation is obtained according to the relative relationship between the keyword discussion volume and the warning keyword discussion volume, and the discussion volume deviation is obtained by processing the ratio operation of the keyword discussion volume and the warning keyword discussion volume; Obtaining a warning delay according to a relative relationship between the keyword appearance time, the warning time and a preset warning interval, wherein the warning delay is obtained by processing a ratio operation of the keyword appearance time and the warning time; The discussion quantity deviation and warning delay are processed to obtain the real-time evaluation value of the warning.

4. The automated public opinion risk early warning application method as claimed in claim 2, characterized in that: The specific method for obtaining the data delay evaluation value based on the acquired initial warning data is as follows: Acquire preset warning data from a preset database, wherein the preset warning data includes a preset Ping value, a preset packet loss rate, a preset I / O waiting time, a preset maximum CPU utilization rate, and a preset maximum memory utilization rate; Obtaining a ping value deviation according to a relative relationship between the ping value and the preset ping value, wherein the ping value deviation is obtained by processing a ratio operation of the ping value and the preset ping value; Obtaining a data packet loss rate deviation according to a relative relationship between the data packet loss rate and a preset data packet loss rate, wherein the data packet loss rate deviation is obtained by processing a ratio operation of the data packet loss rate and the preset data packet loss rate; Obtaining an update coefficient according to a relative relationship between bandwidth and content update frequency, wherein the update coefficient is obtained by processing a ratio operation of bandwidth and content update frequency; A CPU utilization deviation is obtained according to a relative relationship between the CPU utilization and a preset maximum CPU utilization, wherein the CPU utilization deviation is obtained by processing a ratio operation of the CPU utilization and the preset maximum CPU utilization; A memory utilization deviation is obtained according to a relative relationship between the memory utilization and a preset maximum memory utilization, wherein the memory utilization deviation is obtained by processing a ratio operation of the memory utilization and the preset maximum memory utilization; Obtaining an I / O waiting time deviation according to a relative relationship between the I / O waiting time and the preset I / O waiting time, wherein the I / O waiting time deviation is obtained by processing a ratio operation of the I / O waiting time and the preset I / O waiting time; The ping value deviation, packet loss rate deviation, update coefficient, CPU utilization deviation, memory utilization deviation and I / O waiting time deviation are processed to obtain the data delay evaluation value.

5. The automated public opinion risk early warning application method as claimed in claim 4, characterized in that: The specific process of determining whether to perform early warning optimization based on the data delay evaluation value is as follows: Determine whether the data delay evaluation value is less than the preset delay threshold: If the data delay assessment value is less than the preset delay threshold, no early warning optimization will be performed; If the data delay assessment value is not less than the preset delay threshold, early warning optimization is performed; The early warning optimization includes network optimization and resource optimization.

6. The automated public opinion risk early warning application method as claimed in claim 5, characterized in that: The specific process of network optimization is as follows: The network optimization includes bandwidth optimization, content distribution network acceleration, multi-threaded crawling and load balancing; The bandwidth optimization includes increasing network bandwidth and data compression; The content distribution network acceleration includes setting a cache strategy for static content and a cache strategy for dynamic content.

7. The automated public opinion risk early warning application method as claimed in claim 5, characterized in that: The specific process of resource optimization is as follows: The resource optimization includes algorithm reconstruction, multi-thread optimization, use of asynchronous I / O, reduction of CPU frequency and memory management optimization; The memory management optimization includes memory allocation strategy and use of memory pool.

8. The automated public opinion risk early warning application method as claimed in claim 2, characterized in that: The specific method for obtaining the data optimization evaluation value based on the initial warning data and the warning data after warning optimization is as follows: Obtaining a Ping value optimization coefficient according to a relative relationship between the optimized Ping value and the Ping value, wherein the Ping value optimization coefficient is obtained by processing a ratio operation between the optimized Ping value and the Ping value; Obtaining a packet loss rate optimization coefficient according to a relative relationship between the optimized packet loss rate and the packet loss rate, wherein the packet loss rate optimization coefficient is obtained by processing a ratio operation of the optimized packet loss rate and the packet loss rate; Obtaining a bandwidth optimization coefficient according to a relative relationship between the optimized bandwidth and the bandwidth, wherein the bandwidth optimization coefficient is obtained by processing a ratio operation of the optimized bandwidth and the bandwidth; Obtaining a throughput optimization coefficient according to a relative relationship between the optimized throughput, the throughput, and the preset throughput, wherein the throughput optimization coefficient is obtained by processing a ratio calculation between the optimized throughput and the preset throughput and a ratio calculation between the throughput and the preset throughput; The Ping value optimization coefficient, the packet loss rate optimization coefficient, the bandwidth optimization coefficient and the throughput optimization coefficient are processed to obtain a network optimization effective coefficient, which is used to evaluate the effectiveness of network optimization; A CPU utilization optimization coefficient is obtained according to a relative relationship between the optimized CPU utilization deviation and the CPU utilization deviation, wherein the CPU utilization optimization coefficient is obtained by processing a difference between the optimized CPU utilization deviation and the CPU utilization deviation; Obtaining a memory utilization optimization coefficient according to a relative relationship between the optimized memory utilization deviation and the memory utilization deviation, wherein the memory utilization optimization coefficient is obtained by processing a difference between the optimized memory utilization deviation and the memory utilization deviation; Obtaining an I / O waiting time optimization coefficient according to a relative relationship between the optimized I / O waiting time and the I / O waiting time, wherein the I / O waiting time optimization coefficient is obtained by processing a ratio operation of the optimized I / O waiting time and the I / O waiting time; The CPU utilization optimization coefficient, memory utilization optimization coefficient and I / O waiting time optimization coefficient are processed to obtain the resource optimization effective coefficient; The network optimization effectiveness coefficient and resource optimization effectiveness coefficient are processed to obtain the data optimization evaluation value.

9. The automated public opinion risk early warning application method as claimed in claim 8, characterized in that: The specific limiting expression of the data optimization evaluation value is as follows: In the formula, i represents the number of times of public opinion monitoring, i=1,2,...,T, T represents the total number of times of public opinion monitoring, YOU i represents the data optimization evaluation value of the ith public opinion monitoring, WL i represents the network optimization effectiveness coefficient of the ith public opinion monitoring, ZY i It represents the resource optimization efficiency coefficient of the i-th public opinion monitoring.

10. The automated public opinion risk early warning application method as claimed in claim 1, characterized in that: It also includes early warning of public opinion risks. The specific steps are: B1, set the preset media website URL and note website name; B2, set preset keywords and set risk rating; B3, search according to the preset website URL and preset keywords, save the URL corresponding to the preset keyword, and form a risk rating label according to the preset keyword.

Images