Context-Aware Network Text Biased Geographic Information Self-Correction Method

Abstract

The present invention discloses a method, device, storage medium, and computer system for self-correcting biased geographical information of network texts based on context, including: parsing to obtain the geographical information to be corrected in the target sentence of the paragraph, specifically referring to time information and spatial information; screening and analyzing multiple time information and spatial information in the target sentence of the paragraph to perform self-correction of geographical information; screening and judging geographical information from the adjacent sentences in the paragraph context to determine the time information and spatial information that can be self-corrected; and using the precise context to perform self-correction of geographical information. The present invention can effectively utilize the relevant geographical information content in the context information, reduce the deviation of geographical information in the target sentence, make the obtained target geographical information more accurate, be applicable to network texts containing geographical information such as news and social media, and better serve the utilization of network geographical information resources.

Description

Technical Field

[0001] The present invention relates to the fields of artificial intelligence and computational linguistics, and particularly relates to a method, device, storage medium and computer system for self-correcting biased geographical information in network texts based on context. Background Art

[0002] In the Internet era, a vast amount of text data floods various social media, news platforms and forums, and these texts contain rich geographical information. However, with the development of technology and the in-depth application of data, the accuracy and reliability of geographical information in these network texts have become key issues. Geographical information is widely used not only in fields such as navigation, emergency response, and urban planning, but also plays an important role in geographical data analysis, public opinion monitoring, etc. However, the geographical information contained in network texts often exhibits inaccuracies and biases, seriously affecting the effectiveness and reliability of this information in practical applications.

[0003] Specifically, the problems of geographical information deviation in network texts are manifested as inconsistent geographical location descriptions, time point deviations, large differences in information sources, etc. For example, the location information of the "Geographical Information Conference" shows very different descriptions in different social media or news reports, such as "Deqing International Conference Center", "Novotel Hotel", "Deqing Geographical Information Town", etc. Although these place names are close to each other, the actual locations they refer to are different. This difference in expression not only increases the difficulty of automatic parsing of geographical information, but also poses a huge challenge to subsequent applications. Another typical example is the report of a typhoon event. When some news described that "Typhoon Lekima landed in Zhejiang early in the morning", the specific location and time were not accurately described, and there was a large deviation compared with the actual situation. In fact, the accurate location where Typhoon "Lekima" landed was Chengnan Town, Wenling City, and the time was 1:45 in the early morning, while some media reports might generally state "landing in Zhejiang", and this deviation may lead to delays or incorrect judgments in emergency responses.

[0004] These phenomena expose the major challenges faced by geographical information in network texts during extraction and application. Especially in scenarios such as emergency event management and disaster monitoring that require precise geographical information support, inaccurate or biased geographical information may lead to serious consequences. Therefore, how to automatically and accurately correct the geographical information deviation existing in network texts has become an urgent problem to be solved in current research and practical applications. Summary of the Invention

[0005] In view of the above deficiencies in the prior art, the present invention provides a method, apparatus, storage medium, and computer system for self-correcting biased geographical information in network texts based on context, which solves the problem of correction deviation of geographical information (especially time information and spatial location information) contained in network texts, making the geographical information obtained from network texts more accurate.

[0006] In a first aspect, the present application provides a method for self-correcting biased geographical information in network texts based on context. The technical solutions adopted by the method include the following steps:

[0007] S1. Parse and obtain the geographical information element GeoInfo to be self-corrected in the target sentence S of the specified paragraph P. t in the

[0008] S2. Based on multiple pieces of information in the target sentence S t , self-correct the geographical information element GeoInfo according to the in-sentence self-correction algorithm.

[0009] S3. From the context sentences {S k} included in the paragraph P, perform geographical information screening and judgment according to the paragraph self-correction algorithm.

[0010] S4. Use the accurate context of the two layers of in-sentence and paragraph to correct the geographical information element GeoInfo.

[0011] Further, step S1 includes the following sub-steps:

[0012] S1-1. Use a time recognition method or a spatial location recognition method, not limited to specific implementation manners, to recognize the target sentence S of the specified paragraph P t .

[0013] S1-2. Obtain the geographical information element GeoInfo to be self-corrected = {T o , L o}. Wherein, T o is the time information to be self-corrected; L o is the spatial information to be self-corrected.

[0014] Further, the specific method of step S2 includes the following sub-steps:

[0015] S2-1. Use a time recognition method or a spatial location recognition method, not limited to specific implementation manners, to recognize the sentence S where the geographical information element GeoInfo is located t .

[0016] S2-2. Obtain the sequentialized sequence T t of in-sentence time elements = {T ti}, where each time element in the sequence is arranged in the order of its appearance in the text, and i represents the order of the appearance of the time element from the front to the back in the text; and the continuously occurring time element descriptions t 1 t 2 t 3 Need to be additionally recorded as a T tsi , which constitutes the sequential sequence T ts ={T tsi}};

[0017] S2-3. Obtain the sequential sequence L of the in-sentence spatial position elements t ={L ti}}, where each spatial position element in the sequence is arranged in the order of its appearance in the text, and i represents the order of the appearance of the spatial position element from the front to the back in the text; and the continuously occurring spatial position element descriptions l 1 l 2 l 3 Need to be additionally recorded as an L tsi , which constitutes the sequential sequence L ts ={L tsi}};

[0018] S2-4. According to the in-sentence time self-correction function formula:

[0019]

[0020] Perform a correction operation on the time information L to be self-corrected. Among them, T o is the input time information to be self-corrected; T o is the input time information to be self-corrected; T ts [1] represents the first value of the sequential sequence T of the continuously occurring time element descriptions in step S2-2 ts ; T t [1] represents the first value of the sequential sequence T in step S2-2 t ; ft() represents the in-sentence time self-correction function;

[0021] S2-5. According to the in-sentence spatial position self-correction function formula:

[0022]

[0023] Perform a correction operation on the spatial position information L to be self-corrected. Among them, L o is the input spatial position information to be self-corrected; L o is the input spatial position information to be self-corrected; L ts [1] represents the first value of the sequential sequence L of the continuously occurring spatial position element descriptions in step S2-3 ts ; fl() represents the in-sentence spatial position self-correction function.

[0024] Further, the specific method of step S3 includes the following sub-steps:

[0025] S3-1. Sequentially parse the context statements {S k} included in paragraph P to respectively obtain the time element sequentialization sequence T sk = {T skj}, and the time element sequentialization sequence T sks = {T sksj} with consecutive characters. Wherein, S k represents the k-th sentence in paragraph P, and j represents the order of the time element from front to back in sentence k.

[0026] S3-2. Sequentially parse the context statements {S k} included in paragraph P to respectively obtain the spatial position element sequentialization sequence L sk = {L skj}, and the spatial position element sequentialization sequence L sks = {L sksj} with consecutive characters. Wherein, S k represents the k-th sentence in paragraph P, and j represents the order of the spatial position element from front to back in sentence k.

[0027] S3-3. Confirm and find the first time element with consecutive characters in paragraph P, and perform screening according to the following formula:

[0028]

[0029] Wherein, T s0 represents the first time element with consecutive characters, T sks represents the time element sequentialization sequence with consecutive characters in step S3-1, represents from the first sentence of paragraph P to the t-th sentence S t .

[0030] S3-4. Confirm and find, in paragraph P, the spatial position element with consecutive characters that is closest to the spatial position element to be corrected, and perform screening according to the following formula:

[0031]

[0032] Wherein, L s0 represents the first spatial position element with consecutive characters, L sks represents the spatial position element sequentialization sequence with consecutive characters in step S3-2, represents from the first sentence of paragraph P to the t-th sentence S t .

[0033] Further, the specific method of step S4 includes the following sub-steps:

[0034] S4-1. Self-correct the formula according to the time element in the precise context of the context:

[0035]

[0036] Perform correction operations on the time information T to be self-corrected o . Among them, T o is the input time information to be self-corrected; gt() represents the self-correction function of the time element in the precise context of the context; T s0 represents the time element with consecutive first characters in step S3-3; ft(T o ) represents the result of the intra-sentence time self-correction function in step S2-4;

[0037] S4-2. Self-correct the formula according to the spatial position element in the precise context of the context:

[0038]

[0039] Perform correction operations on the spatial position information L to be self-corrected o . Among them, L o is the input spatial position information to be self-corrected; gl() represents the self-correction function of the spatial position element in the precise context of the context; L s0 represents the spatial position element with consecutive first characters in step S3-4; fl(L o ) represents the result of the intra-sentence spatial position self-correction function in step S2-5;

[0040] S4-3. Obtain the context self-correction result {gt(T o ), gl(L o )} of the geographical information element GeoInfo = {T o , L o}.

[0041] In the second aspect, the present application provides a biased geographical information self-correction device for network text based on context, which is used to implement the above method. The device includes the following modules:

[0042] Parsing module: used to parse and obtain the time and spatial information to be self-corrected;

[0043] Intra-sentence self-correction module: self-correct the time and spatial information based on the intra-sentence context;

[0044] Paragraph self-correction module: perform information screening and correction in combination with the context of the paragraph context;

[0045] Context correction module: Utilize the double-layer context within sentences and paragraphs to accurately correct geographical information elements;

[0046] The above modules work together to enable the device to efficiently achieve precise correction of geographical information in network texts, and are applicable to the processing of various text data containing geographical information.

[0047] In a third aspect, the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a computer system, the computer system can implement the above self-correction method. The storage medium can be in various forms such as a hard disk, flash memory, optical disc, cloud storage, etc., to ensure that the program can be executed in multiple environments.

[0048] In a fourth aspect, the present invention also provides a computer system for performing self-correction of geographical information based on context. The computer system includes a processor, a memory, and an interface module for data input and result output. By invoking the program stored on the storage medium, the computer system can parse and correct geographical information, effectively eliminating the deviation of geographical information in the text.

[0049] The beneficial effects of the present invention are as follows:

[0050] 1. By combining the context to perform self-correction on the biased geographical information in network texts, the present invention can improve the accuracy of geographical information and significantly alleviate the problems caused by inconsistent or inaccurate description of geographical information. This is of great significance for ensuring the reliability of geographical information in applications such as navigation, emergency response, and urban planning.

[0051] 2. By using the context for multi-level analysis, the present invention can automatically identify and correct the biased geographical information according to the overall semantics and background information of the text, thereby ensuring that the final geographical information is more in line with the actual situation. This can effectively reduce misjudgment and misuse caused by information deviation, especially in the decision-making and management of major events.

[0052] 3. The present invention can flexibly adapt to different types of texts and scenarios. Through an adaptive correction mechanism, it can reduce the deviation of geographical information in complex or dynamic texts. For example, in texts with real-time nature and complex backgrounds such as social media and news reports, the present invention can effectively improve the accuracy of geographical information.

[0053] 4. The interpretability of the present invention provides users with a transparent understanding of the correction process. By explaining the basis for geographical information correction, users can clearly understand how the correction process comprehensively considers various factors of context and context, thereby enhancing their trust and dependence on the system. At the same time, this feature can provide scientific basis and decision-making support for researchers, helping them better understand the sources of deviation and correction mechanisms of geographical information in texts.

[0054] 5. The present invention can be widely applied to a variety of network text data sources, automatically correct the biased geographical information in the text, effectively reduce the workload of manual correction, thereby improving the efficiency and accuracy of data processing. This will greatly promote the deep integration and application of geographical information systems (GIS) and natural language processing technologies in multiple fields. Through these beneficial effects, the present invention provides an innovative method for the efficient and accurate correction of network text geographical information, helps to improve the accuracy and practicality of geographical information, and promotes the intelligent development of data applications and decision support systems in related fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 is a schematic diagram of the overall process of this method;

[0056] Figure 2 is an example of network text for the implementation of this method;

[0057] Figure 3 is a schematic diagram of the self-correction computer system structure proposed by the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0058] The following describes the specific implementation manners of the present invention to facilitate those skilled in the art of this technology to understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation manners. For those of ordinary skill in the art of this technology, as long as various changes are within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions created using the concept of the present invention are within the scope of protection.

[0059] As Figure 1 shown in the technical route, in an embodiment of the present invention, it includes the following steps: target statement parsing (parsing and obtaining the geographical information elements to be self-corrected in the specified paragraph, including time and space information); in-sentence self-correction (using a self-correction algorithm to self-correct the geographical information elements based on the in-sentence context); paragraph self-correction (screening relevant geographical information from the paragraph context and making a screening judgment in combination with the context within the paragraph); context correction (correcting the geographical information elements through the in-sentence and paragraph-level context to ensure higher accuracy). Taking Figure 2 a network news report as an implementation example, as Figure 2 shown, it includes the following sub-steps:

[0060] S11. Use the extractEntities parsing method in the HanLP open source package to parse the time information and spatial location of the target statement "Recently, in Shilong Village, Liangping Township, Wushan County, Chongqing City, villager Xiang Changjun was busy in his own field."

[0061] S12, obtaining the geographic information element GeoInfo to be self-calibrated = {"recently", "own field"}, wherein "recently" is the time information to be self-calibrated; "own field" is the spatial information to be self-calibrated.

[0062] S21. Use the extractEntities method in the HanLP open source package to identify the sentence where the geographic information element GeoInfo is located: "Recently, in Shilong Village, Liangping Township, Wushan County, Chongqing, villager Xiang Changjun was busy in his own field."

[0063] S22. Obtain the sequential sequence T of the time elements in the sentence t ={"recently"}; and the time elements of the characters are continuous to describe the sequence T ts ={""};

[0064] S23, obtain the sequence L of spatial position elements in the sentence t ={"Chongqing City","Wushan County","Liangping Township","Shilong Village","Zijiatianli"},; and the character continuous spatial position element describes the sequential sequence L ts ={"Shilong Village, Liangping Township, Wushan County, Chongqing"};

[0065] S24, perform a correction operation on the time information to be self-corrected "recent days". ts is empty, so ft(T o )=T t ={“recently”};

[0066] S25, the spatial position information to be self-corrected "my own field" is corrected. ts is not empty, so fl(L o )=L ts [1] = {“Shilong Village, Liangping Township, Wushan County, Chongqing”};

[0067] S31, parse the context sentences S contained in paragraph P in sequence 1 Chongqing, January 24th "The weather has been very good these past two days, so our whole family has been out harvesting radishes in the fields. We have harvested more than 10,000 kilograms, and there are still about 20,000 kilograms in the fields." 2 "Recently, in Shilong Village, Liangping Township, Wushan County, Chongqing, villager Xiang Changjun was busy in his own fields.", S 3 "Xiang Changjun was busy working while talking about how the big-headed radishes were growing well this season. After being pickled, the green radishes turned into "gold nuggets". ", won T s1 ={"January","24th"}, T s1s ={"January 24"}, Ts2 = {"Recently"}, T s2s = {""}, T s3 = {"One season"}, T s3s = {""};

[0068] S32. Analyze the context statements S included in paragraph P in sequence 1 : "Chongqing, January 24th. 'The weather has been great these two days. Our whole family went out to harvest turnips in the field. We've already harvested over 10,000 jin, and there are still about 20,000 jin left in the field.'", S 2 : "Recently, in Shilong Village, Liangping Township, Wushan County, Chongqing, villager Xiang Changjun has been very busy in his own field.", S 3 : "While working, Xiang Changjun muttered that the turnips in this season are growing very well. These green turnips, after being pickled, have become 'golden lumps'.", obtaining L s1 = {"Chongqing", "field"}, L s1s = {""}, L s2 = {"Chongqing City", "Wushan County", "Liangping Township", "Shilong Village", "his own field"}, L s2s = {"Shilong Village, Liangping Township, Wushan County, Chongqing City"}, L s3 = {""}, L s3s = {""};

[0069] S33. Confirm and find the first consecutive time element T in the paragraph s0 = {"January 24th"};

[0070] S34. Confirm and find the spatial position element L that is closest to the position element of the space to be corrected and has consecutive characters in the paragraph s0 = {"Shilong Village, Liangping Township, Wushan County, Chongqing City"};

[0071] S41. Perform a correction operation on the time information "Recently" to be self-corrected because Therefore, gt(T o ) = T s0 = {"January 24th"};

[0072] S42. Perform a correction operation on the spatial position information "his own field" to be self-corrected because , therefore, gl(L o ) = L s0 = {"Shilong Village, Liangping Township, Wushan County, Chongqing City"};

[0073] S43. Obtain the context self-correction result {"January 24th", "Shilong Village, Liangping Township, Wushan County, Chongqing City"} for the geographical information elements GeoInfo = {"recently", "in my own field"}.

[0074] In addition, this application also proposes a device for self-correcting biased geographical information in network texts based on context, which is used to implement the above method and includes the following modules:

[0075] Parsing module: Used to parse the target statement and extract the time and space information to be self-corrected;

[0076] Self-correction module: Use the information within the sentence to self-correct the geographical information elements and generate a sequential sequence of time and space positions;

[0077] Paragraph self-correction module: Screen the relevant time and space information in the paragraph context, compare and confirm the optimal correction elements;

[0078] Context correction module: Generate a correction result based on the in-sentence and paragraph context to improve the information accuracy.

[0079] Meanwhile, this application proposes a storage medium for self-correcting biased geographical information in network texts based on context. The storage medium stores program code that can be executed in a computer system and is suitable for implementing the method for self-correcting biased geographical information in network texts based on context of the present invention. The storage medium can be a non-volatile storage device such as a hard disk, SSD, optical disc, etc.

[0080] In addition, this application proposes a computer system for self-correcting biased geographical information in network texts based on context. As Figure 3 shown, this system automatically implements the method for self-correcting biased geographical information in network texts based on context of the present invention by calling the program in the storage medium and includes the following components: Processor M3: Used to execute the method for self-correcting biased geographical information in network texts based on context; Memory M2: Store and read the time and space information of the text to be corrected; Input interface module M1: Receive the text information to be corrected; Output interface module M4: Output the corrected geographical information result for further processing or display.

[0081] In summary, the present invention can utilize other clues in the text to judge the accuracy of geographical information through an adaptive correction mechanism. This method can not only identify biased geographical information in the text, but also perform adaptive correction based on context clues, thereby improving the accuracy and reliability of geographical information. Through multi-level semantic understanding of the text and cross-text information comparison, the self-correction method based on context can significantly reduce errors and enhance the effectiveness of geographical information in network texts in various applications. This provides an important technical means for future automated processing of geographical information in massive network texts, and also opens up new directions for research and applications in related fields.

Claims

1. A method for self-correcting biased geographic information in network text based on context, characterized in that: The following steps are involved: S1. Parse and obtain the target sentence S of the specified paragraph P t The geographic information element GeoInfo to be self-corrected; S2, based on the target sentence S t The plurality of information in the sentence are self-corrected according to the self-correction algorithm within the sentence, and the geographic information element GeoInfo is self-corrected; step S2 includes the following sub-steps: S2-1, using the time identification method or the spatial location identification method, the statement S t To identify; S2-2. Obtaining the sequential sequence T of the time elements in the sentence t ={T ti }, where each time element in the sequence is arranged in the order of its appearance in the text, i represents the order in which the time element appears from the beginning to the end in the text; and the time element description t1t2t3 with continuous characters needs to be recorded as an additional T tsi , which constitutes a sequential sequence T ts ={T tsi }; S2-3. Obtain the ordered sequence L of the spatial position elements in the sentence t ={L ti }, where each spatial position element in the sequence is arranged in the order of its appearance in the text, i represents the order in which the spatial position element appears from the front to the back in the text; and the spatial position element descriptions l1l2l3 with consecutive characters need to be recorded as an additional L tsi , which constitutes the sequential sequence L ts ={L tsi }; S2-4, according to the sentence time self-correction function formula: The time information to be self-corrected T o Perform correction operation, where T o is the input time information to be self-calibrated; T ts [1] represents the sequential sequence T of the continuous time element description in step S2-2 ts The first value of T t [1] represents the serialization sequence T in step S2-2 t The first value of ; ft() represents the intra-sentence time self-correction function; S2-5, according to the sentence spatial position self-correction function formula: The spatial position information to be self-corrected L o Perform correction operation, where L o is the input spatial position information to be self-corrected; L ts [1] represents the sequential sequence L of the continuous spatial position element description in step S2-3 ts The first value of; fl() represents the spatial position self-correction function within the sentence; S3, from the context sentence {S k }, according to the paragraph self-correction algorithm, the geographic information screening and judgment are performed; step S3 includes the following sub-steps: S3-1. Analyze the context sentences contained in paragraph P in sequence {S k }, and obtain the time element sequence T of each statement respectively sk ={T skj }, and the time element sequence T of continuous characters sks ={T sksj }, where S k represents the kth sentence in paragraph P, and j represents the order in which the time element appears from the beginning to the end in sentence k; S3-2, parse the context sentences contained in paragraph P in sequence {S k }, and obtain the ordered sequence L of the spatial position elements of each sentence sk ={L skj }, and the character continuous spatial position element sequence L sks ={L sksj }, where S k represents the kth sentence in paragraph P, and j represents the order in which the spatial position element appears from the front to the back in sentence k; S3-3. Confirm and find the first continuous time element of characters in paragraph P, and filter it according to the following formula: Among them, T s0 Indicates the time element of the first character continuation, T sks represents the character continuous time element sequence in step S3-1, Indicates the time information from the first sentence of paragraph P to the tth sentence S to be self-corrected t ; S3-4. Confirm and find the spatial position element with the closest distance to the spatial position element to be corrected in paragraph P, with continuous characters, and select it according to the following formula: Among them, L s0 Indicates the continuous spatial position element of the first character, L sks represents the ordered sequence of continuous spatial position elements of characters in step S3-2, It represents the space from the first sentence of paragraph P to the tth sentence S where the spatial position information to be self-corrected is located. t ; S4, using the precise context of the sentence and paragraph to correct the geographic information element GeoInfo; step S4 includes the following sub-steps: S4-1. Self-correction formula based on the time element in the precise context: The time information to be self-corrected T o Perform correction operation, where T o is the input time information to be self-corrected; gt() represents the time element self-correction function in the context precision context; T s0 represents the time element of the first character continuation in step S3-3; ft(T o ) represents the result of the intra-sentence time self-correction function in step S2-4; S4-2. Self-correction formula based on spatial position elements in the precise context: The spatial position information to be self-corrected L o Perform correction operation, where L o is the input spatial position information to be self-corrected; gl() represents the self-correction function of the spatial position element in the context precision context; L s0 represents the spatial position element of the first character in step S3-4; fl(L o ) represents the result of the intra-sentence spatial position self-correction function in step S2-5; S4-3, obtain geographic information element GeoInfo = {T o ,L o }'s contextual self-correction result {gt(T o ),gl(L o )}.

2. According to the context-based network text biased geographic information self-correction method of claim 1, it is characterized by: Step S1 includes the following sub-steps: S1-1. Use the time recognition method or the spatial position recognition method to identify the target sentence S in the specified paragraph P. t To identify; S1-2, obtain the geographic information element GeoInfo to be self-corrected = {T o ,L o }, where T o is the time information to be self-corrected; L o is the spatial information to be self-corrected.

3. A device for self-correcting biased geographic information in network text based on context, characterized in that: Includes the following modules: A parsing module, used for parsing a target sentence of a specified paragraph to obtain geographic information elements to be self-corrected; The intra-sentence self-correction module corrects the geographic information elements of the target sentence based on the intra-sentence self-correction algorithm; it includes: S2-1, using the time identification method or the spatial location identification method, the statement S t To identify; S2-2. Obtaining the sequential sequence T of the time elements in the sentence t ={T ti }, where each time element in the sequence is arranged in the order of its appearance in the text, i represents the order in which the time element appears from the beginning to the end in the text; and the time element description t1t2t3 with continuous characters needs to be recorded as an additional T tsi , which constitutes a sequential sequence T ts ={T tsi }; S2-3. Obtain the ordered sequence L of the spatial position elements in the sentence t ={L ti }, where each spatial position element in the sequence is arranged in the order of its appearance in the text, i represents the order in which the spatial position element appears from the front to the back in the text; and the spatial position element descriptions l1l2l3 with consecutive characters need to be recorded as an additional L tsi , which constitutes the sequential sequence L ts ={L tsi }; S2-4, according to the sentence time self-correction function formula: The time information to be self-corrected T o Perform correction operation, where T o is the input time information to be self-calibrated; T ts [1] represents the sequential sequence T of the continuous time element description in step S2-2 ts The first value of T t [1] represents the serialization sequence T in step S2-2 t The first value of ; ft() represents the intra-sentence time self-correction function; S2-5, according to the sentence spatial position self-correction function formula: The spatial position information to be self-corrected L o Perform correction operation, where L o is the input spatial position information to be self-corrected; L ts [1] represents the sequential sequence L of the continuous spatial position element description in step S2-3 ts The first value of; fl() represents the spatial position self-correction function within the sentence; The paragraph self-correction module extracts key information from the context of the paragraph and uses the paragraph self-correction algorithm to filter and correct geographic information; including: S3-1. Analyze the context sentences {S k }, and obtain the time element sequence T of each statement respectively sk ={T skj }, and the time element sequence T of continuous characters sks ={T sksj }, where S k represents the kth sentence in paragraph P, and j represents the order in which the time element appears from the beginning to the end in sentence k; S3-2, parse the context sentences contained in paragraph P in sequence {S k }, and obtain the ordered sequence L of the spatial position elements of each sentence sk ={L skj }, and the character continuous spatial position element sequence L sks ={L sksj }, where S k represents the kth sentence in paragraph P, and j represents the order in which the spatial position element appears from the front to the back in sentence k; S3-3. Confirm and find the first continuous time element of characters in paragraph P, and filter it according to the following formula: Among them, T s0 Indicates the time element of the first character continuation, T sks represents the character continuous time element sequence in step S3-1, Indicates the time information from the first sentence of paragraph P to the tth sentence S to be self-corrected t ; S3-4. Confirm and find the spatial position element with the closest distance to the spatial position element to be corrected in paragraph P, with continuous characters, and select it according to the following formula: Among them, L s0 Indicates the continuous spatial position element of the first character, L sks represents the ordered sequence of continuous spatial position elements of characters in step S3-2, It represents the space from the first sentence of paragraph P to the tth sentence S where the spatial position information to be self-corrected is located. t ; The contextual self-correction module performs final correction of geographic information elements based on the contextual context within the sentence and paragraph level; it includes: S4-1. Self-correction formula based on the time element in the precise context: The time information to be self-corrected T o Perform correction operation, where T o is the input time information to be self-corrected; gt() represents the time element self-correction function in the context precision context; T s0 represents the time element of the first character continuation in step S3-3; ft(T o ) represents the result of the intra-sentence time self-correction function in step S2-4; S4-2. Self-correction formula based on spatial position elements in the precise context: The spatial position information to be self-corrected L o Perform correction operation, where L o is the input spatial position information to be self-corrected; gl() represents the self-correction function of the spatial position element in the context precision context; L S0 represents the spatial position element of the first character in step S3-4; fl(L o ) represents the result of the intra-sentence spatial position self-correction function in step S2-5; S4-3, obtain geographic information element GeoInfo = {T o ,L o }'s contextual self-correction result {gt(T o ),gl(L o )}.

4. A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, enables the processor to execute any one of the context-based self-correction methods for biased geographic information in network texts according to claims 1-2.

5. A computer system comprising: A processor, configured to execute instructions stored in a storage medium to implement any one of claims 1-2 of the method for self-correction of biased geographic information in network text based on context; A memory, used to store the network text to be self-corrected and the intermediate results generated during the calculation process; The input / output interface is used to receive a specified text paragraph or geographic information and output corrected geographic information elements.

Images