A method and device for correcting errors in meteorological disaster warning information

Abstract

The present invention relates to the technical field of meteorological disaster warning, and proposes a method and device for correcting meteorological disaster warning information errors, including obtaining a first meteorological keyword based on meteorological data, obtaining a second meteorological keyword based on a meteorological disaster warning information text, determining the weight of the keyword, and then calculating the similarity between the first meteorological keyword and the second meteorological keyword based on the weight. When setting the weight for the keyword, the comprehensive impact of the combined occurrence of different disaster types is fully considered, and it can be judged whether the prediction of the meteorological data by the meteorological disaster warning information text to be detected is accurate, avoiding the warning information errors caused by only checking the form problem, and realizing automatic, efficient and accurate meteorological disaster warning information error correction; through the above technical scheme, the problem of low accuracy in identifying errors in warning information in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of meteorological disaster warning technology, and in particular to a method and device for correcting errors in meteorological disaster warning information. Background Art

[0002] The accuracy and timeliness of meteorological disaster warning information plays a vital role in mitigating natural disasters and protecting people's lives and property. At a time when extreme weather events are becoming more frequent, an efficient and accurate meteorological disaster warning system can provide the government and the public with valuable response time to take necessary preventive measures, thereby effectively reducing the losses caused by meteorological disasters.

[0003] However, existing technologies face multiple limitations when processing meteorological disaster warning information, mainly in terms of low efficiency and high error rate. First, many existing systems rely on manual analysis and proofreading, which is time-consuming and labor-intensive. Existing verification systems often run slowly when processing and analyzing large amounts of meteorological data, and cannot achieve real-time or near-real-time warning releases, making it difficult to meet the needs of rapid response in emergencies. Second, due to the lack of advanced automated detection tools, existing systems are not very accurate in identifying errors in warning information, which may mislead the public and increase disaster risks.

[0004] Therefore, there is an urgent need to conduct research on error correction of meteorological disaster warning information in order to achieve more accurate and efficient meteorological disaster warnings. Summary of the invention

[0005] The present invention provides a method and device for correcting errors in meteorological disaster warning information, which solves the problem that the existing method for correcting errors in meteorological disaster warning information cannot accurately and efficiently perform error correction on warning systems in different regions.

[0006] The technical solution of the present invention is as follows:

[0007] A method for correcting errors in meteorological disaster warning information comprises the following steps:

[0008] Step S1, obtaining meteorological data to be forecasted and meteorological disaster warning information text corresponding to the meteorological data;

[0009] Step S2, inputting the meteorological data into a trained meteorological keyword prediction model, and outputting a first meteorological keyword;

[0010] Step S3, inputting the meteorological disaster warning information text into a trained meteorological keyword extraction model, and outputting a second meteorological keyword;

[0011] Step S4, assigning a basic weight to each keyword according to the degree of influence of each keyword on the warning information; the keywords included in the first meteorological keyword and the second meteorological keyword are: disaster type, warning level, affected area, preventive measures, call for action, impact time, disaster intensity, issuing agency, professional terminology, risk warning, contact information, and expected impact;

[0012] Step S5, determining the interaction coefficient affecting the basic weight according to the type of disaster type;

[0013] Step S6, calculating the similarity between the first meteorological keyword and the second meteorological keyword according to the basic weight and the interaction coefficient; if the similarity is less than a preset threshold, outputting that there is a forecast error in the meteorological disaster warning information, otherwise there is no forecast error;

[0014] Step S7, performing spelling detection, punctuation detection, and logic detection on the meteorological disaster warning information text by using text proofreading software; outputting the detection result to determine whether there is a formal error or not;

[0015] Step S8, when there is no forecast error and no form error at the same time, then there is no error in the meteorological disaster warning information text, otherwise it is output that there is an error, and at the same time outputs the error type and error location.

[0016] Furthermore, the training method of the meteorological keyword prediction model is:

[0017] Step S21, obtaining historical meteorological data and meteorological disaster warning information text corresponding to the meteorological data;

[0018] Step S22, preprocessing the meteorological data, wherein the preprocessing includes data cleaning, standardization, and normalization;

[0019] Step S23, preprocessing the meteorological disaster warning information text, the preprocessing including cleaning, word segmentation, removal of stop words, and keyword marking;

[0020] Step S24, taking the preprocessed meteorological data as input and the keywords as output, training the neural network model to obtain a trained meteorological keyword prediction model.

[0021] Furthermore, the training method of the meteorological keyword extraction model is:

[0022] Step S31, obtaining historical meteorological disaster warning information text;

[0023] Step S32, preprocessing the meteorological disaster warning information text, the preprocessing including cleaning, word segmentation, removal of stop words, and keyword marking;

[0024] Step S33, taking the pre-processed meteorological disaster warning information text as input and the keywords as output, training the natural language processing model NLP to obtain a trained meteorological keyword extraction model.

[0025] Furthermore, the disaster types include: typhoon / hurricane, heavy rain, drought, blizzard / snow disaster, cold wave, high temperature / heat wave, low temperature / cold current, hail, sandstorm, lightning, strong wind / storm, haze, flood, mountain torrent / mudslide, tsunami, and red tide.

[0026] Furthermore, in step S4, the method for setting the basic weight of the keyword is:

[0027] Set the disaster type, warning level, impact time, impact area, preventive measures, and call for action to high weight A;

[0028] Set the disaster intensity, expected impact, risk warning, and contact information to medium weight B;

[0029] Set the publishing organization and professional terminology to a low weight C;

[0030] Among them, A>B>C.

[0031] Furthermore, in step S5, the interaction coefficient affecting the basic weight is determined according to the type of disaster, and the specific method is as follows:

[0032] Step S51, data collection: collecting historical meteorological disaster event data, including the time, location, type and severity of the disaster;

[0033] Step S52, event analysis: obtaining the actual warning level, impact time, disaster intensity, and risk warning information of each disaster type when each disaster type occurs alone; obtaining the actual warning level, impact time, disaster intensity, and risk warning information of each disaster type when different disaster types occur simultaneously or continuously;

[0034] Step S53, statistical modeling: using statistical methods to collect statistics on the warning level, impact time, disaster intensity, and risk warning information of each disaster type to determine the interaction relationship between different disasters;

[0035] Step S54, risk assessment: assess the impact of different disaster type combinations on human safety, property, infrastructure and environment, and adjust the interaction relationship between different disasters based on the impact results.

[0036] Furthermore, in step S6, the similarity calculation formula is:

[0037] S=(S1+S3+S4+S5)×A+(S 11 +S 12 )×B+(S8+S9)×C+(S2×A+(S6+S7+S 10 )×B)×n;

[0038] Among them, S is similarity, S1 is disaster type, S2 is warning level, S3 is affected area, S4 is preventive measures, S5 is call for action, S6 is impact time, S7 is disaster intensity, S8 is issuing agency, S9 is professional terminology, S 10 As a risk warning, S 11 For contact information, S 12 is the estimated impact, A is high weight, B is medium weight, C is low weight, and n is the interaction coefficient.

[0039] Furthermore, the text proofreading software is Grammarly.

[0040] A meteorological disaster warning information error correction device, using a meteorological disaster warning information error correction method as described in any one of the above items, comprises the following modules:

[0041] A data acquisition module, which acquires the meteorological data to be forecasted and the meteorological disaster warning information text corresponding to the meteorological data;

[0042] A keyword acquisition module, connected to the data acquisition module, is used to input the meteorological data into a trained meteorological keyword prediction model to output a first meteorological keyword; input the meteorological disaster warning information text into a trained meteorological keyword extraction model to output a second meteorological keyword;

[0043] A weight determination module: connected to the keyword acquisition module, used to assign a basic weight to each keyword according to the degree of influence of each keyword on the warning information; determine the interaction coefficient that affects the basic weight according to the type of disaster type; the keywords included in the first meteorological keyword and the second meteorological keyword are: disaster type, warning level, affected area, preventive measures, call for action, impact time, disaster intensity, issuing agency, professional terminology, risk warning, contact information, and expected impact;

[0044] A forecast error checking module: connected to the weight determination module, used to calculate the similarity between the first meteorological keyword and the second meteorological keyword according to the basic weight and the interaction coefficient; if the similarity is less than a preset threshold, outputting that there is a forecast error in the meteorological disaster warning information, otherwise there is no forecast error;

[0045] A formal error checking module is connected to the forecast error checking module, and performs spelling detection, punctuation detection, and logic detection on the meteorological disaster warning information text through text proofreading software; and outputs the detection result whether there is a formal error or not;

[0046] Output module: connected to the forecast error checking module and the form error checking module. When there is no forecast error and form error at the same time, there is no error in the meteorological disaster warning information text. Otherwise, it outputs that there is an error, and outputs the error type and error location at the same time.

[0047] The working principle and beneficial effects of the present invention are:

[0048] 1. In the present invention, a first meteorological keyword is obtained according to meteorological data, a second meteorological keyword is obtained according to a meteorological disaster warning information text, the weight of the keyword is determined, and then the similarity between the first meteorological keyword and the second meteorological keyword is calculated according to the weight, so as to judge whether the prediction of the meteorological disaster warning information text to be detected for the current meteorological data is accurate, thereby avoiding the error of the warning information caused by only checking the form problem, and realizing automatic, efficient and accurate error correction of the meteorological disaster warning information;

[0049] 2. When setting weights for keywords, the combined impact of different disaster types occurring together is fully considered, further improving the accuracy of early warning information. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0051] Figure 1 This is a flow chart of the meteorological disaster warning information error correction method of the present invention;

[0052] Figure 2 The present invention is a flow chart of a training method for a meteorological keyword prediction model;

[0053] Figure 3 The flowchart of the training method of the meteorological keyword extraction model of the present invention;

[0054] Figure 4 A schematic diagram of a device for correcting errors in meteorological disaster warning information according to the present invention;

[0055] Figure 5 This is a structural block diagram of an electronic device provided for this embodiment 3. DETAILED DESCRIPTION

[0056] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] Example 1

[0058] like Figure 1 As shown, this embodiment proposes a method for correcting errors in meteorological disaster warning information, comprising the following steps:

[0059] Step S1, obtaining the meteorological data to be forecasted and the meteorological disaster warning information text corresponding to the meteorological data.

[0060] The role of step S1 is to provide basic data and text information for the implementation of the meteorological disaster warning information error correction method. The meteorological data to be forecasted is used to predict meteorological keywords to determine whether the disaster type, forecast level, etc. in the meteorological disaster warning information text are accurate. Avoiding warning information errors caused by only checking formal problems is an effective means to achieve automatic, efficient and accurate meteorological disaster warning information error correction.

[0061] In step S2 and step S3, these data and texts will be used to predict and extract meteorological keywords, which is a key step for further error correction.

[0062] Step S2, inputting the meteorological data into a trained meteorological keyword prediction model, and outputting a first meteorological keyword.

[0063] like Figure 2 As shown in the figure, the training method of the meteorological keyword prediction model is:

[0064] Step S21, obtaining historical meteorological data and meteorological disaster warning information text corresponding to the meteorological data;

[0065] Step S22, preprocessing the meteorological data, wherein the preprocessing includes data cleaning, standardization, and normalization;

[0066] Step S23, preprocessing the meteorological disaster warning information text, the preprocessing includes cleaning, word segmentation, removing stop words, and keyword marking;

[0067] Step S24, taking the preprocessed meteorological data as input and the keywords as output, training the neural network model to obtain a trained meteorological keyword prediction model.

[0068] Step S3, inputting the meteorological disaster warning information text into a trained meteorological keyword extraction model, and outputting a second meteorological keyword.

[0069] like Figure 3 As shown in Figure 1, the training method of the meteorological keyword extraction model is:

[0070] Step S31, obtaining historical meteorological disaster warning information text;

[0071] Step S32, preprocessing the meteorological disaster warning information text, the preprocessing including cleaning, word segmentation, removal of stop words, and keyword marking;

[0072] Step S33, taking the pre-processed meteorological disaster warning information text as input and the keywords as output, training the natural language processing model NLP to obtain a trained meteorological keyword extraction model.

[0073] In the process of acquiring keywords, the present invention adopts different models according to different input types.

[0074] The input of the meteorological keyword prediction model is meteorological data, including structured data, such as temperature, humidity, wind speed, air pressure and other values. Therefore, the preprocessing methods include data cleaning, standardization and normalization, and the processing is relatively simple. The model structure selects a machine learning model that is easy to process structured data, such as a neural network model.

[0075] The input of the meteorological keyword extraction model is the meteorological disaster warning information text, which is unstructured data. During preprocessing, complex text preprocessing such as text cleaning, word segmentation, stop word removal, part-of-speech tagging, and named entity recognition is required. Since features need to be extracted from the text, special processing methods are selected, such as bag-of-words model, TF-IDF, Word Embeddings, etc. The model processes complex text information and needs to consider contextual information, so the natural language processing model NLP is selected for processing. Its training data is meteorological disaster warning information text and keyword labels, and the text data needs to have rich language expressions.

[0076] Step S4, assigning a basic weight to each keyword according to the degree of influence of each keyword on the warning information; the keywords included in the first meteorological keyword and the second meteorological keyword are: disaster type, warning level, affected area, preventive measures, call for action, impact time, disaster intensity, issuing agency, professional terminology, risk warning, contact information, and expected impact. In other embodiments, it is not limited to the above keywords.

[0077] Disaster types include: typhoon / hurricane, rainstorm, drought, snowstorm / snowstorm, cold wave, high temperature / heat wave, low temperature / cold wave, hail, sandstorm, lightning, strong wind / storm, haze, flood, mountain torrent / mudslide, tsunami, red tide. In other embodiments, it is not limited to the above disaster types.

[0078] The method for setting the basic weight of keywords is:

[0079] Set the disaster type, warning level, impact time, impact area, preventive measures, and call for action to high weight A;

[0080] Set the disaster intensity, expected impact, risk warning, and contact information to medium weight B;

[0081] Set the publishing organization and professional terminology to a low weight C;

[0082] There is a hierarchical relationship between weights, i.e. A>B>C, which ensures that in the similarity calculation, keywords with higher weights have a greater impact on the results.

[0083] Disaster type is given a high weight: because the type of disaster directly determines the nature of the warning and the preventive measures that the public needs to take;

[0084] Warning levels are given high weight: Warning levels are a key indicator of the severity of disasters and are crucial to public response;

[0085] Impact time is given medium to high weights: Knowing when a disaster will occur can help the public and relevant departments prepare in time;

[0086] Impact areas are given high weight: Identifying the affected areas helps identify specific locations where action is needed;

[0087] Preventive measures are given high weight: Specific preventive measures are one of the most important parts of warning information, directly guiding the public on how to act;

[0088] Calls to action are given high weight: Calls to action are a key part in motivating the public to take necessary actions;

[0089] Disaster intensity is assigned a medium weight: Disaster intensity information helps assess the likely extent of damage and the response measures required.

[0090] Weights in expected impacts: Understanding the possible impacts can help the public and relevant departments assess risks and take appropriate preventive measures;

[0091] Risk warnings are assigned a medium weight: Risk warnings help the public understand possible secondary disasters so that they can take more comprehensive preventive measures.

[0092] Contact information is given medium weight: In an emergency, it is crucial to provide emergency contact information;

[0093] The issuing agency is given a low weight: The authority of the issuing agency can increase the credibility of the warning information.

[0094] Professional terms are given low weight: For professionals, professional terms help to understand the disaster situation more accurately.

[0095] Warning levels include: blue warning, yellow warning, orange warning, and red warning.

[0096] The impact time includes: warning start time, duration, and end time.

[0097] The affected area includes: the affected city, region, province, country or geographic coordinates.

[0098] Disaster intensity includes: wind level, rainfall, temperature range, snow amount, etc.

[0099] The expected impacts include: traffic disruption, crop damage, impact on residents’ lives, infrastructure damage, etc.

[0100] Prevention measures include: recommended actions, safety guidelines, shelter information, supply preparation, and more.

[0101] Issuing agencies include: Meteorological Bureau, emergency management department, government agencies, etc.

[0102] Contact information includes: rescue phone, police phone, official consultation channels, etc.

[0103] Professional terms include: meteorological terms such as cyclone, front, trough, etc.

[0104] Risk warnings include: possible secondary disasters, such as landslides, urban flooding, etc.

[0105] Calls to action include: calling on the public to pay attention to safety, reduce going out, pay attention to official information, etc.

[0106] These keywords help to construct a clear and accurate meteorological disaster warning message, ensuring that the information is conveyed quickly and effectively. In the actual warning message, the use of keywords will be adjusted according to the specific situation and disaster characteristics.

[0107] Step S4 is a key link in the method for correcting errors in meteorological disaster warning information. Its purpose is to assign weights to keywords and consider the combined impact of disaster types when they occur together to improve the accuracy of warning information. Through this step, errors in warning information can be more effectively identified and corrected, thereby improving the efficiency and effectiveness of the warning system.

[0108] Step S5, determining the interaction coefficient affecting the basic weight according to the type of disaster.

[0109] Considering the combined impact of different disaster types, the basic weight is adjusted by determining the interaction coefficient. This reflects that when multiple disasters occur at the same time, they may affect each other, thus affecting the accuracy of early warning information. The specific method is:

[0110] Step S51, data collection: collecting historical meteorological disaster event data, including the time, location, type and severity of the disaster;

[0111] Step S52, event analysis: obtaining the actual warning level, impact time, disaster intensity, and risk warning information of each disaster type when each disaster type occurs alone; obtaining the actual warning level, impact time, disaster intensity, and risk warning information of each disaster type when different disaster types occur simultaneously or continuously;

[0112] Step S53, statistical modeling: using statistical methods to collect statistics on the warning level, impact time, disaster intensity, and risk warning information of each disaster type to determine the interaction relationship between different disasters;

[0113] Step S54, risk assessment: assess the impact of different disaster type combinations on human safety, property, infrastructure and environment, and adjust the interaction relationship between different disasters based on the impact results.

[0114] When multiple meteorological disaster warnings are issued at the same time, the risk level may increase significantly, because different disaster types may affect each other and increase the severity of the disaster. The following warning combinations may lead to an increase in risk level:

[0115] Heavy rain warnings and flood alerts: Heavy rainfall may cause river and lake levels to rise rapidly, causing flooding;

[0116] Typhoon warning and heavy rain warning: The strong winds and heavy rainfall brought by typhoons may cause serious flooding and damage;

[0117] Heavy rain warning and flash flood warning: Heavy rainfall in mountainous areas can easily cause flash floods and mudslides;

[0118] Cold Wave Warning and High Wind Warning: Cold temperatures combined with strong winds may cause severe wind chill effects, increasing the risk of hypothermia;

[0119] High temperature warning and drought warning: continued high temperatures and drought may lead to water shortages and increase the pressure on crops and vegetation;

[0120] Dust storm warning and high wind warning: strong winds carry dust and sand, reducing visibility and causing serious impacts on human health and traffic;

[0121] Hail warning and high wind warning: Hail accompanied by strong winds may pose a double threat to crops, buildings and personal safety;

[0122] Blizzard warning and cold wave warning: Heavy snowfall and sudden drop in temperature may cause severe freezing, affecting transportation and life;

[0123] Lightning warning and high wind warning: Lightning activity accompanied by strong winds may increase the risk of outdoor work and activities;

[0124] Haze warning and low temperature warning: Haze weather may be more difficult to dissipate under low temperature conditions, causing long-term impacts on human health;

[0125] Typhoon warning and sea wave warning: Typhoons may cause rough seas, posing a threat to coastal areas and maritime navigation safety.

[0126] In these combined situations, the potential impacts of disasters may overlap, leading to more extensive damage and more complex emergency response needs. Therefore, when multiple warnings appear at the same time, relevant departments and the public need to take more stringent preventive measures and make adequate emergency preparations.

[0127] In step S53, a statistical method is used to collect statistics on the warning level, impact time, disaster intensity, and risk warning information of each disaster type to determine the interaction relationship between different disasters, which can be obtained through the following steps:

[0128] Step S531, data collection and cleaning: collect historical meteorological disaster data, including disaster type, warning level, impact time, disaster intensity and risk warning; clean the data and remove incomplete or inaccurate records;

[0129] Step S532, feature extraction: extracting features related to disasters, such as the frequency, duration, intensity distribution, etc. of disasters;

[0130] Step S533, disaster type classification: classify the data according to the disaster type and analyze the characteristics of each disaster;

[0131] Step S534, association rule learning: applying an association rule learning algorithm (such as an Apriori algorithm) to discover the correlation between different disaster types;

[0132] Step S535, time series analysis: performing time series analysis on the time data of disaster occurrence to identify periodicity and trend;

[0133] Step S536, spatial analysis: using a geographic information system (GIS) to analyze the spatial distribution and interrelationships of disasters;

[0134] Step S537, statistical analysis: Apply statistical tests (such as chi-square test, correlation analysis) to evaluate the significance between different disasters; construct statistical models (such as regression analysis, survival analysis) to evaluate the interaction between disasters.

[0135] It is also possible to conduct risk assessment on the models used in the above analysis process, simulate the potential impact of different disaster combinations on people, property and the environment, or interpret and adjust the statistical results in combination with expert knowledge.

[0136] The following table is the interaction relationship table obtained in this embodiment:

[0137]

[0138] In other embodiments, the interaction relationship is not limited to the above.

[0139] Step S6, calculating the similarity between the first meteorological keyword and the second meteorological keyword according to the basic weight and the interaction coefficient; if the similarity is less than a preset threshold, outputting that there is a forecast error in the meteorological disaster warning information, otherwise there is no forecast error.

[0140] The calculation formula of the similarity is:

[0141] S=(S1+S3+S4+S5)×A+(S 11 +S 12 )×B+(S8+S9)×C+(S2×A+(S6+S7+S 10 )×B)×n;

[0142] Among them, S is similarity, S1 is disaster type, S2 is warning level, S3 is affected area, S4 is preventive measures, S5 is call for action, S6 is impact time, S7 is disaster intensity, S8 is issuing agency, S9 is professional terminology, S 10 As a risk warning, S 11 For contact information, S 12 is the estimated impact, A is high weight, B is medium weight, C is low weight, and n is the interaction coefficient.

[0143] This method emphasizes keywords through positive weighting, which can effectively measure the accuracy of warning information while keeping the calculation simple and intuitive. Each keyword (S1 to S12) is assigned a different weight (A, B, C) according to its importance, and then added to obtain an intuitive and easy-to-interpret similarity composite score. This method simplifies the calculation process, avoids the problems that may be caused by division operations, and comprehensively reflects the degree of match between warning information and meteorological data through cumulative effects.

[0144] In step S6, the similarity between the first meteorological keyword obtained from the meteorological data and the second meteorological keyword extracted from the meteorological disaster warning information text is calculated based on the basic weight and interaction coefficient of each keyword determined in the previous step, and the weight (high weight A, medium weight B, low weight C) and interaction coefficient (n) of each keyword are applied to the similarity calculation using a specific calculation formula, and the calculated similarity is compared with a preset threshold. This threshold is a standard for judging whether the similarity of two keyword sets is within an acceptable range: if the similarity is less than the preset threshold, step S6 will output a result indicating that there is a forecast error in the meteorological disaster warning information, indicating that the keywords in the warning information text do not match the keywords predicted by the actual meteorological data, and need to be reviewed and corrected; if the similarity is not less than the preset threshold, it is considered that there is no forecast error, that is, the meteorological disaster warning information text is accurate in predicting the current meteorological data.

[0145] By considering the keyword weights and interaction coefficients, step S6 can more accurately evaluate the accuracy of the warning information, thereby improving the reliability of the meteorological disaster warning system. Step S6 is part of the automated error correction process, which helps reduce the need for manual inspection, improve efficiency, and reduce errors caused by human factors.

[0146] Through step S6, the present invention can automatically, efficiently and accurately perform error correction on meteorological disaster warning information, ensure the accuracy and timeliness of the warning information, and is of great significance to reducing natural disasters and protecting people's lives and property.

[0147] Step S7, performing spelling detection, punctuation detection, and logic detection on the meteorological disaster warning information text through text proofreading software; outputting the detection result to indicate whether there is a formal error or not.

[0148] The text of the meteorological disaster warning information is comprehensively checked using text proofreading software, including spelling, punctuation, grammar, and possible logical consistency detection. The text proofreading software in this embodiment is Grammarly, which is a powerful tool that can detect and correct a variety of errors in the text.

[0149] Spelling check: Text proofreading software checks every word in a text, identifying and marking any spelling errors or suspected errors;

[0150] Punctuation detection: The software analyzes the punctuation marks used in the text to ensure that they are correct and conform to grammatical rules;

[0151] Logical testing: Logical testing may include checking the logical coherence, sentence fluency, and whether there are semantic contradictions or unreasonableness in the text.

[0152] If the text proofreading software finds any formal errors, it will output these errors, including the error type and specific location, for correction; if the text proofreading software does not find any formal errors, it will output a result indicating that the meteorological disaster warning information text is correct in form.

[0153] The result of step S7 will be combined with the forecast error check result of step S6 to determine whether there are any errors in the meteorological disaster warning information text, whether they are prediction errors in content or errors in form.

[0154] Through step S7, the present invention can ensure that the meteorological disaster warning information text is not only accurate in content but also conforms to standards in form before being released, thereby improving the professionalism, readability and reliability of the information.

[0155] Step S8, when there is no forecast error and no form error at the same time, then there is no error in the meteorological disaster warning information text, otherwise it is output that there is an error, and at the same time outputs the error type and error location.

[0156] In this embodiment, a first meteorological keyword is obtained based on meteorological data, a second meteorological keyword is obtained based on a meteorological disaster warning information text, the weight of the keyword is determined, and then the similarity between the first meteorological keyword and the second meteorological keyword is calculated based on the weight. When setting the weight for the keyword, the comprehensive impact of the combined occurrence of different disaster types is fully considered. It can be judged whether the meteorological disaster warning information text to be detected is accurate in its prediction of the current meteorological data, avoiding warning information errors caused by only checking formal problems, and realizing automatic, efficient and accurate correction of meteorological disaster warning information errors. Through the above technical scheme, the problem of low accuracy in identifying errors in warning information in the prior art is solved.

[0157] Example 2

[0158] like Figure 4 As shown, based on the same concept as the above-mentioned embodiment 1, this embodiment also proposes a meteorological disaster warning information error correction device, using a meteorological disaster warning information error correction method as described in any one of the embodiments 1, including the following modules:

[0159] A data acquisition module, which acquires the meteorological data to be forecasted and the meteorological disaster warning information text corresponding to the meteorological data;

[0160] A keyword acquisition module, connected to the data acquisition module, is used to input the meteorological data into a trained meteorological keyword prediction model to output a first meteorological keyword; input the meteorological disaster warning information text into a trained meteorological keyword extraction model to output a second meteorological keyword;

[0161] A weight determination module: connected to the keyword acquisition module, used to assign a basic weight to each keyword according to the degree of influence of each keyword on the warning information; determine the interaction coefficient that affects the basic weight according to the type of disaster type; the keywords included in the first meteorological keyword and the second meteorological keyword are: disaster type, warning level, affected area, preventive measures, call for action, impact time, disaster intensity, issuing agency, professional terminology, risk warning, contact information, and expected impact;

[0162] A forecast error checking module: connected to the weight determination module, used to calculate the similarity between the first meteorological keyword and the second meteorological keyword according to the basic weight and the interaction coefficient; if the similarity is less than a preset threshold, outputting that there is a forecast error in the meteorological disaster warning information, otherwise there is no forecast error;

[0163] A formal error checking module is connected to the forecast error checking module, and performs spelling detection, punctuation detection, and logic detection on the meteorological disaster warning information text through text proofreading software; and outputs the detection result whether there is a formal error or not;

[0164] Output module: connected to the forecast error checking module and the form error checking module. When there is no forecast error and form error at the same time, there is no error in the meteorological disaster warning information text. Otherwise, it outputs that there is an error, and outputs the error type and error location at the same time.

[0165] In this embodiment, the data acquisition module is responsible for collecting the meteorological data to be forecasted and the corresponding meteorological disaster warning information text, and providing the necessary input data for subsequent error correction; the keyword acquisition module uses the trained meteorological keyword prediction model and the meteorological keyword extraction model to extract key information from the meteorological data and the warning information text respectively, and obtains two groups of keywords; the weight determination module assigns a basic weight to each keyword according to its influence on the warning information, and determines the interaction coefficient according to the type of disaster type. These weights and coefficients will be used for subsequent similarity calculations; the forecast error check module uses the basic weight and interaction coefficient to calculate the similarity of the two groups of keywords, and judges whether there is a forecast error in the warning information by comparing the similarity with the preset threshold; the formal error check module uses text proofreading software (such as Grammarly) to perform spelling, punctuation and logic detection on the warning information text to identify formal errors; the output module comprehensively judges whether there is an error in the warning information text based on the inspection results of the two. If there is neither a forecast error nor a formal error, the output text is error-free; if there is any error, the output has an error, and the type and specific location of the error are indicated.

[0166] The meteorological disaster warning information error correction device of Example 2 realizes automated and efficient error detection of meteorological disaster warning information through the collaborative work of these modules, improves the accuracy and reliability of warning information, and helps reduce potential risks caused by erroneous information.

[0167] Example 3

[0168] like Figure 5 As shown, the present invention further provides an electronic device 100, and the electronic device 100 includes:

[0169] Processor 101 and memory 102;

[0170] The processor 101 is used to execute the steps of the meteorological disaster warning information error correction method as described in any one of Example 1 by calling the program or instruction stored in the memory 102.

[0171] Example 4

[0172] A computer-readable storage medium, the computer-readable storage medium comprising computer program instructions, the computer program instructions causing a computer to execute the steps of the meteorological disaster warning information error correction method as described in any one of Example 1.

[0173] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0174] It should be noted that the terms used in the present invention are only for describing specific embodiments, rather than limiting the scope of the present invention. As shown in the present specification, unless the context clearly indicates an exception, the words "one", "a", "a kind of" and / or "the" do not specifically refer to the singular, but may also include the plural. The terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method or device. In the absence of more restrictions, the elements defined by the sentence "include one..." do not exclude the presence of other identical elements in the process, method or device including the elements.

[0175] It should also be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. Unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", etc. should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be a connection between the two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0176] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for correcting errors in meteorological disaster warning information, characterized in that: The steps include: Step S1, obtaining the meteorological data to be forecasted and the meteorological disaster warning information text corresponding to the meteorological data; Step S2, inputting the meteorological data into a trained meteorological keyword prediction model, and outputting a first meteorological keyword; Step S3, inputting the meteorological disaster warning information text into a trained meteorological keyword extraction model, and outputting a second meteorological keyword; Step S4, assigning a basic weight to each keyword according to the degree of influence of each keyword on the warning information; the keywords included in the first meteorological keyword and the second meteorological keyword are: disaster type, warning level, affected area, preventive measures, call for action, impact time, disaster intensity, issuing agency, professional terminology, risk warning, contact information, and expected impact; the method for setting the basic weight of the keyword is: Set the disaster type, warning level, impact time, impact area, preventive measures, and call for action to high weight A; Set the disaster intensity, expected impact, risk warning, and contact information to medium weight B; Set the publishing organization and professional terminology to a low weight C; Among them, A>B>C; Step S5, determining the interaction coefficient affecting the basic weight according to the type of disaster type; Step S6, calculating the similarity between the first meteorological keyword and the second meteorological keyword according to the basic weight and the interaction coefficient; if the similarity is less than a preset threshold, outputting that there is a forecast error in the meteorological disaster warning information, otherwise there is no forecast error; the calculation formula of the similarity is: S=(S1+S3+S4+S5)×A+(S 11 +S 12 )×B+(S8+S9)×C+(S2×A+(S6+S7+S 10 )×B)×n; Among them, S is similarity, S1 is disaster type, S2 is warning level, S3 is affected area, S4 is preventive measures, S5 is call for action, S6 is impact time, S7 is disaster intensity, S8 is issuing agency, S9 is professional terminology, S 10 As a risk warning, S 11 For contact information, S 12 is the estimated impact, A is high weight, B is medium weight, C is low weight, and n is the interaction coefficient; Step S7, performing spelling detection, punctuation detection, and logic detection on the meteorological disaster warning information text by using text proofreading software; outputting the detection result to determine whether there is a formal error or not; Step S8, when there is no forecast error and no form error at the same time, then there is no error in the meteorological disaster warning information text, otherwise it is output that there is an error, and at the same time outputs the error type and error location.

2. A method for correcting errors in meteorological disaster warning information according to claim 1, characterized in that: The training method of the meteorological keyword prediction model is: Step S21, obtaining historical meteorological data and meteorological disaster warning information text corresponding to the meteorological data; Step S22, preprocessing the meteorological data, wherein the preprocessing includes data cleaning, standardization, and normalization; Step S23, preprocessing the meteorological disaster warning information text, the preprocessing including cleaning, word segmentation, removal of stop words, and keyword marking; Step S24, taking the preprocessed meteorological data as input and the keywords as output, training the neural network model to obtain a trained meteorological keyword prediction model.

3. A method for correcting errors in meteorological disaster warning information according to claim 1, characterized in that: The training method of the meteorological keyword extraction model is: Step S31, obtaining historical meteorological disaster warning information text; Step S32, preprocessing the meteorological disaster warning information text, the preprocessing including cleaning, word segmentation, removal of stop words, and keyword marking; Step S33, taking the pre-processed meteorological disaster warning information text as input and the keywords as output, training the natural language processing model NLP to obtain a trained meteorological keyword extraction model.

4. A method for correcting errors in meteorological disaster warning information according to claim 1, characterized in that: The disaster types include: typhoon / hurricane, heavy rain, drought, blizzard / snow disaster, cold wave, high temperature / heat wave, low temperature / cold current, hail, sandstorm, lightning, strong wind / storm, haze, flood, mountain torrent / mudslide, tsunami, and red tide.

5. A method for correcting errors in meteorological disaster warning information according to claim 1, characterized in that: In step S5, the interaction coefficient affecting the basic weight is determined according to the type of disaster type. The specific method is: Step S51, data collection: collecting historical meteorological disaster event data, including the time, location, type and severity of the disaster; Step S52, event analysis: obtaining the actual warning level, impact time, disaster intensity, and risk warning information of each disaster type when each disaster type occurs alone; obtaining the actual warning level, impact time, disaster intensity, and risk warning information of each disaster type when different disaster types occur simultaneously or continuously; Step S53, statistical modeling: using statistical methods to collect statistics on the warning level, impact time, disaster intensity, and risk warning information of each disaster type to determine the interaction relationship between different disasters; Step S54, risk assessment: assess the impact of different disaster type combinations on human safety, property, infrastructure and environment, and adjust the interaction relationship between different disasters based on the impact results.

6. A method for correcting errors in meteorological disaster warning information according to claim 1, characterized in that: The text proofreading software is Grammarly.

7. A meteorological disaster warning information error correction device, using a meteorological disaster warning information error correction method according to any one of claims 1 to 6, characterized in that: Includes the following modules: A data acquisition module, which acquires the meteorological data to be forecasted and the meteorological disaster warning information text corresponding to the meteorological data; A keyword acquisition module, connected to the data acquisition module, is used to input the meteorological data into a trained meteorological keyword prediction model to output a first meteorological keyword; input the meteorological disaster warning information text into a trained meteorological keyword extraction model to output a second meteorological keyword; Weight determination module: connected to the keyword acquisition module, used to assign a basic weight to each keyword according to the degree of influence of each keyword on the warning information; According to the types of disasters, the interaction coefficients that affect the basic weights are determined; The keywords included in the first meteorological keyword and the second meteorological keyword are: disaster type, warning level, affected area, preventive measures, call for action, impact time, disaster intensity, issuing agency, professional terminology, risk warning, contact information, and expected impact; the method for setting the basic weight of the keyword is: Set the disaster type, warning level, impact time, impact area, preventive measures, and call for action to high weight A; Set the disaster intensity, expected impact, risk warning, and contact information to medium weight B; Set the publishing organization and professional terminology to a low weight C; Among them, A>B>C; A forecast error checking module: connected to the weight determination module, used to calculate the similarity between the first meteorological keyword and the second meteorological keyword according to the basic weight and the interaction coefficient; if the similarity is less than a preset threshold, outputting that there is a forecast error in the meteorological disaster warning information, otherwise there is no forecast error; The formal error checking module is connected to the forecast error checking module, and performs spelling detection, punctuation detection, and logic detection on the meteorological disaster warning information text through text proofreading software; outputs the detection result to determine whether there is a formal error or not; the similarity calculation formula is: S=(S1+S3+S4+S5)×A+(S 11 +S 12 )×B+(S8+S9)×C+(S2×A+(S6+S7+S 10 )×B)×n; Among them, S is similarity, S1 is disaster type, S2 is warning level, S3 is affected area, S4 is preventive measures, S5 is call for action, S6 is impact time, S7 is disaster intensity, S8 is issuing agency, S9 is professional terminology, S 10 As a risk warning, S 11 For contact information, S 12 is the estimated impact, A is high weight, B is medium weight, C is low weight, and n is the interaction coefficient; Output module: connected to the forecast error checking module and the form error checking module. When there is no forecast error and form error at the same time, there is no error in the meteorological disaster warning information text. Otherwise, it outputs that there is an error, and outputs the error type and error location at the same time.

Images