A method for measuring the quality of geographic knowledge maps based on life cycle theory

Abstract

A method for measuring the quality of a geographic knowledge graph based on the life cycle theory divides the life cycle of a geographic knowledge graph into data stage, production stage, content stage, application stage and exchange stage. According to the definition of the geographic knowledge graph life cycle, the quality of the geographic knowledge graph is comprehensively measured by the quality of the data stage, the quality of the production stage, the quality of the content stage, the quality of the application stage and the quality of the exchange stage. The quality of the data stage is measured using the reliability of the knowledge source, the spatiotemporal coverage of the knowledge source and the reliability of the knowledge medium. The quality of the production stage is measured using the performance index of the conversion method and the confidence index of the conversion method. The quality of the content stage is measured using the two indicators of knowledge correctness and knowledge error. The quality of the application stage is measured using the two indicators of knowledge integrity and knowledge representativeness. The quality of the exchange stage is measured using the two indicators of interpretability and exchangeability. The quality of each link in the life cycle of the geographic knowledge graph can be measured, and the quality of the geographic knowledge graph can be comprehensively reflected with high accuracy.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and in particular to a method for measuring the quality of a geographic knowledge graph based on life cycle theory. Background Art

[0002] With the development of information and communication technology, massive spatiotemporal information continues to emerge, showing the characteristics of multi-source heterogeneity, which poses a challenge to the organization and mining of spatiotemporal knowledge contained in it. To this end, the geographic knowledge graph, which combines the general knowledge graph and the characteristics of spatiotemporal information, came into being. The geographic knowledge graph is a graph-structured spatiotemporal knowledge organization method based on nodes and relationships. It uses RDF triples to store geographic concepts and their relationships, geographic objects, and aggregated spatiotemporal information. This spatiotemporal information aggregated around geographic entities can effectively express geographic characteristics such as time, space, state, process, and evolution, and supports applications such as content association, compound query, collaborative computing, and logical reasoning of spatiotemporal knowledge, and is gradually becoming the driving core of geoscientific intelligent applications. The effects of the above applications depend to a large extent on the quality of the geographic knowledge graph. Therefore, users need to measure the quality of the geographic knowledge graph before using it.

[0003] However, the industry currently lacks a systematic method for measuring the quality of geographic knowledge graphs, which is reflected in the direct use of quality measurement methods for general knowledge graphs to measure the quality of geographic knowledge graphs, leading to two major problems: First, the measurement task is only aimed at a specific goal, and the indicators and calculation methods used are scattered, and no complete quality measurement framework has been formed; second, the measurement method of general knowledge graphs does not take into account the spatiotemporal characteristics of geographic knowledge and cannot correctly reflect the quality of geographic knowledge graphs.

[0004] Therefore, technicians in this field are committed to developing a method for measuring the quality of geographic knowledge graphs based on life cycle theory to address the deficiencies of the above-mentioned prior art. Summary of the invention

[0005] In view of the above-mentioned defects of the prior art, the technical problem to be solved by the present invention is that the currently existing public technology directly uses the quality measurement method of the general knowledge graph to measure the quality of the geographic knowledge graph, resulting in the measurement task being only for a specific target, and the indicators and calculation methods used are scattered, and no complete quality measurement framework has been formed; and the measurement method of the general knowledge graph does not take into account the spatiotemporal characteristics of geographical knowledge, and cannot correctly reflect the defects of the quality of the geographic knowledge graph.

[0006] To achieve the above-mentioned purpose, the present invention provides a method for measuring the quality of a geographic knowledge graph based on the life cycle theory, including dividing the life cycle of a geographic knowledge graph into a data stage, a production stage, a content stage, an application stage, and an exchange stage, and defining the quality of a geographic knowledge graph according to the life cycle of the geographic knowledge graph. GeoKG Data stage quality ins , Quality in the production stage mth , Content stage quality con , Application stage quality app and exchange phase quality acc Comprehensive determination; specifically includes the following steps,

[0007] Step 1: Calculate the data quality ins ; Quality of the data stage ins The quality of the input data of the geographic knowledge graph to be measured includes the following steps:

[0008] Step 1-1, calculate the reliability of the knowledge source; the knowledge source reliability indicator Indicator prv The data stage quality is measured using the reliability of the knowledge source for constructing the geographic knowledge graph to be measured, and the higher the reliability of the knowledge source for constructing the geographic knowledge graph to be measured, the higher the data stage quality;

[0009] The calculation of the reliability of the knowledge source specifically includes the following steps:

[0010] Step 1-1-1: Calculate the hierarchical weight prv of a knowledge source weight , the information level of knowledge source prv information ;

[0011] The information level prv of the knowledge source information The calculation formula is:

[0012]

[0013] in,

[0014] provider level is the registration level or reputation level index of the knowledge source;

[0015] provider number is the contribution index of the knowledge source;

[0016] provider distance is the spatial deviation distance index between the knowledge source and the provided knowledge;

[0017] Furthermore, the registration level or reputation level index of the knowledge source provider level Use the following calculation formula to get:

[0018]

[0019] in,

[0020] Level(provider) is the registration level or reputation level of the knowledge source;

[0021] MAX_LEVEL is the maximum registration level or reputation level in all knowledge sources.

[0022] The higher the registration level or reputation level of the knowledge source, the higher the registration level or reputation level index provider level The bigger;

[0023] Furthermore, the contribution index of the knowledge source provider number Use the following calculation formula to get:

[0024]

[0025] in,

[0026] number(provider) is the amount of knowledge contributed by the knowledge source;

[0027] AVERAGE_NUMBER is the average amount of knowledge contributed by all knowledge sources;

[0028] Therefore, the more knowledge the knowledge source contributes, the higher the contribution index of the knowledge source provider number The bigger;

[0029] Furthermore, the spatial deviation distance index between the knowledge source and the provided knowledge provider distance Use the following calculation formula to get:

[0030]

[0031] distance(providercity-object) is the Euclidean distance between the knowledge source and the provided knowledge in geographic space;

[0032] The smaller the Euclidean distance between the knowledge source and the provided knowledge in geographic space, the higher the spatial deviation distance index provider distance The bigger;

[0033] Furthermore, the hierarchical weight prv of the knowledge source weight Determined by provider tier;

[0034] In a specific embodiment of the present invention, if the provider level is a world or national authoritative data source, the level weight is 1.00;

[0035] In a specific embodiment of the present invention, the provider level is a provincial or regional high-confidence data source, and the level weight is 0.75;

[0036] In a specific implementation of the present invention, the provider level is a general public data source, and the level weight is 0.50;

[0037] In a specific embodiment of the present invention, the provider level is a personal data source, and the level weight is 0.25;

[0038] Step 1-1-2: Calculate the source reliability Prv of one piece of knowledge among the n pieces of knowledge in the geographic knowledge graph to be measured i ; The source reliability of the piece of knowledge Prv i The calculation formula is:

[0039] Prv i =prv weight ×prv information

[0040] in,

[0041] prv weight is the hierarchical weight of the knowledge source;

[0042] prv information is the information level of the knowledge source;

[0043] Step 1-1-3: Calculate the source reliability indicator prv , the knowledge source reliability indicator Indicator prv The calculation formula is:

[0044]

[0045] in,

[0046] The geographic knowledge graph to be determined is constructed from n pieces of knowledge;

[0047] Prv i The reliability of the source of one piece of knowledge;

[0048] Step 1-2: Calculate the knowledge source spatiotemporal coverage indicator crg ; The knowledge source spatiotemporal coverage indicator Indicatorcrg The data stage quality is measured using the spatial and temporal restriction degree of the knowledge source for constructing the geographic knowledge graph to be measured. The higher the spatiotemporal coverage of the knowledge source for constructing the geographic knowledge graph to be measured, the higher the data stage quality; specifically comprising the following steps:

[0049] Step 1-2-1: Calculate the spatial coverage crg of one piece of knowledge among the n pieces of knowledge constituting the geographic knowledge graph to be measured spatial , time coverage crg temporal ;

[0050] Furthermore, the spatial coverage crg spatial Use the following calculation formula to get:

[0051]

[0052] in,

[0053] num spatial is the number of spatial knowledge triples in the geographic knowledge graph to be determined;

[0054] num sp_limited is the number of spatial knowledge triples with spatial area restrictions in the geographic knowledge graph to be determined;

[0055] The fewer spatial area restrictions there are in the geographic knowledge graph to be measured, the greater the spatial coverage crg spatial The bigger;

[0056] Furthermore, the time coverage crg temporal Use the following calculation formula to get:

[0057]

[0058] in,

[0059] num temporal is the number of knowledge triples in the geographic knowledge graph to be determined;

[0060] num tp_limited is the number of spatial knowledge triples with time range restrictions in the geographic knowledge graph to be determined;

[0061] The fewer time range restrictions there are in the geographic knowledge graph to be measured, the greater the time coverage crg temporal The bigger;

[0062] Step 1-2-2: Calculate the spatiotemporal coverage of a piece of knowledge source Crg i ; The calculation formula is:

[0063] Cr i =crg spatial ×crg temporal

[0064] in,

[0065] crg spatial is the spatial coverage, which measures the coverage of the knowledge source in space;

[0066] crg temporal is the temporal coverage, which measures the temporal coverage of the knowledge source;

[0067] Step 1-2-3: Calculate the knowledge source spatiotemporal coverage indicator crg ; The calculation formula is:

[0068]

[0069] in,

[0070] Cr i is the spatiotemporal coverage of the knowledge source of one piece of knowledge among the n pieces of knowledge constituting the geographic knowledge graph to be measured;

[0071] Step 1-3: Calculate the knowledge medium reliability indicator med ; The knowledge medium reliability indicator Indicator med The data stage quality is measured using the reliability of the transmission medium that carries the knowledge for constructing the geographic knowledge graph to be measured. The higher the reliability of the knowledge medium for constructing the geographic knowledge graph to be measured, the higher the data stage quality. Specifically, the following steps are included:

[0072] Step 1-3-1, calculate the knowledge medium suitability med of a piece of knowledge in the said generation determination geographic knowledge graph suitability , media loss type ;

[0073] Furthermore, the knowledge medium suitability med suitability Use the following calculation formula to get:

[0074]

[0075] in,

[0076] The geographic knowledge graph to be determined is constructed from l types of knowledge;

[0077] st i the suitability of the type of medium for one of the categories of knowledge;

[0078] If such knowledge is transmitted through a suitable medium, then the st i The value is 1.00;

[0079] If such knowledge is transmitted through an inappropriate medium, the st i The value is 0.50;

[0080] Furthermore, the media loss med type , determined according to the type of knowledge;

[0081] The knowledge type is structured knowledge, the medium loss is lossless, and the medium lossiness is 1.00;

[0082] The knowledge type is semi-structured knowledge, the media loss is lossy, and the media lossiness is 0.75;

[0083] The knowledge type is unstructured knowledge, the media loss is ambiguous, and the media loss is 0.50;

[0084] Step 1-3-2: Calculate the media reliability Med of a piece of knowledge i The formula is

[0085] Med i =med suitability ×med type

[0086] in,

[0087] med suitability is the suitability of the medium of knowledge, which determines whether the knowledge is transmitted through the most appropriate medium;

[0088] med type The medium lossiness is to determine whether the medium used by the knowledge causes transmission loss to the knowledge;

[0089] Step 1-3-3: Calculate the knowledge medium reliability indicator med ; The knowledge medium reliability indicator Indicator med The calculation formula is:

[0090]

[0091] in,

[0092] Med i is the medium reliability of one piece of knowledge among the n pieces of knowledge constituting the geographic knowledge graph to be measured;

[0093] Step 1-4: Knowledge source reliability indicator obtained based on steps 1-1 to 1-3 prv、Knowledge source spatiotemporal coverage indicator crg 、Knowledge medium reliability indicator med The three indicators comprehensively measure the quality of the data to be measured. ins The formula is

[0094]

[0095] In the formula,

[0096] Indicator prv Reliability of knowledge sources;

[0097] Indicator crg is the spatial and temporal coverage of knowledge sources;

[0098] Indicator med Reliability as a medium of knowledge;

[0099] Step 2: Calculate the quality of the production stage mth ; Quality of the production stage mth The method is to determine the quality of the input data when it is converted into the geographic knowledge graph to be determined, which specifically includes the following steps:

[0100] Step 2-1: Determine the performance of the conversion method performance , conversion method confidence mth type ;

[0101] Furthermore, the conversion method performance mth performance The F1 value in the field of natural language processing can be used for calculation:

[0102]

[0103] in,

[0104] P is the accuracy of the conversion result: P = number of correctly converted data / total number of converted data;

[0105] R is the recall of the conversion result: R = number of correctly converted data / number of data that should be converted;

[0106] Furthermore, the conversion method confidence mtht ype This can be determined using the test level of the conversion result:

[0107] mth type =1-∝

[0108] in,

[0109] ∝ is the test level or significance level of the result;

[0110] Step 2-2: Calculate the method reliability Mth of a conversion method i ; The calculation formula is:

[0111] M i =mth performance ×mth type

[0112] in,

[0113] mth performance is the conversion method performance;

[0114] mtht ype is the conversion method confidence;

[0115] Step 2-3: Determine the production quality of the geographic knowledge graph to be determined mth , the calculation formula is:

[0116]

[0117] in,

[0118] M i reliability of a method for converting the input data into one of the conversion methods for the geographic knowledge graph to be determined using m conversion methods;

[0119] Step 3: Calculate the quality of the content stage con ; Quality of the content stage con The method is to determine the quality of the knowledge contained in the geographic knowledge graph to be determined; specifically, the method comprises the following steps:

[0120] Step 3-1: Calculate the knowledge logic correctness crr of a piece of knowledge in the geographic knowledge graph to be measured loqic 、Correctness of knowledge content content ;

[0121] For the logical correctness of the knowledge crr logic The following calculation formula can be used:

[0122]

[0123] in,

[0124] num wl is the number of knowledge in the geographic knowledge graph to be determined that conflicts with the ontology model;

[0125] n is the number of all knowledge in the geographic knowledge graph to be measured;

[0126] The less the knowledge in the geographic knowledge graph to be measured that conflicts with the ontology model, the more correct the knowledge logic is. logic The bigger;

[0127] The correctness of the knowledge content content The following calculation formula can be used:

[0128]

[0129] in,

[0130] num wt is the amount of knowledge in the geographic knowledge graph to be measured that is inconsistent with the actual situation;

[0131] n is the number of all knowledge in the geographic knowledge graph to be measured;

[0132] The less the knowledge in the geographic knowledge graph to be measured is inconsistent with the actual situation, the more correct the knowledge content is. content The bigger;

[0133] Step 3-2: Calculate the knowledge correctness Crr of a piece of knowledge i ; The calculation formula is:

[0134] Crr i =crr logic ×crr content

[0135] in,

[0136] crr loqic It is the logical correctness of knowledge;

[0137] crr content It is the correctness of knowledge content;

[0138] Step 3-3: Calculate the knowledge correctness indicator of the geographic knowledge graph to be measured crr ; The knowledge correctness indicator Indicator crr The correctness of the knowledge contained in the geographic knowledge graph to be measured is used to measure the quality of the content stage. The higher the correctness of the knowledge in the geographic knowledge graph to be measured, the higher the quality of the content stage. The knowledge correctness indicator Indicator crr The calculation formula is

[0139]

[0140] in,

[0141] Crr iis the knowledge correctness of one piece of knowledge among the n pieces of knowledge constituting the geographic knowledge graph to be measured;

[0142] Step 3-4: Calculate the category error cndt of the geographic knowledge graph knowledge to be measured ype , spatial error of knowledge cnd spatial , the time error of knowledge cnd temporal ;

[0143] Furthermore, the category error cndt ype The following calculation formula can be used:

[0144]

[0145] in,

[0146] num wtt is the number of knowledge with incorrect categories in the geographic knowledge graph to be measured;

[0147] n is the number of all knowledge in the geographic knowledge graph to be measured;

[0148] The smaller the number of knowledge with category errors in the geographic knowledge graph to be measured, the smaller the category error cnd type The bigger;

[0149] Furthermore, the spatial error cnd spatial The following calculation formula can be used:

[0150]

[0151] in,

[0152] err_spatial i is the number of erroneous knowledge in the geographic knowledge graph to be measured at a certain spatial granularity;

[0153] n is the number of all knowledge in the geographic knowledge graph to be measured, and the spatial granularity and its number p can be set with reference to different spatial division methods such as administrative divisions, spatial resolution, and scale;

[0154] The smaller the number of erroneous knowledge in the geographic knowledge graph to be measured in different spatial granularities, the smaller the spatial error cnd spatial The bigger;

[0155] Furthermore, the time error cnd temporal The following calculation formula can be used:

[0156]

[0157] in,

[0158] err_temporal i is the number of erroneous knowledge in the geographic knowledge graph to be measured at a certain time granularity;

[0159] n is the number of all knowledge in the geographic knowledge graph to be measured, and the time granularity and its number q can be set in years, months, days, hours, minutes, seconds, etc.;

[0160] The smaller the number of erroneous knowledge in the geographic knowledge graph to be measured in different time granularities, the smaller the time error cnd temporal The bigger;

[0161] Step 3-5: Calculate the knowledge error indicator Indicator of the geographic knowledge graph knowledge to be measured cnd ; The knowledge error indicator Indicator cnd The content stage quality is measured using the error degree of the knowledge contained in the geographic knowledge graph to be measured. The lower the knowledge error of the geographic knowledge graph to be measured, the higher the content stage quality. The knowledge error indicator Indicator cnd The calculation formula is:

[0162] Indicator cnd =cnd type ×cnd spatial ×cnd temporal

[0163] in,

[0164] cnd type is the category error of the n pieces of knowledge constituting the geographic knowledge graph to be measured;

[0165] cnd spatial is the spatial error of the n pieces of knowledge constituting the geographic knowledge graph to be measured;

[0166] cnd temporal is the time error of the n pieces of knowledge constituting the geographic knowledge graph to be measured;

[0167] Step 3-6: Indicator of knowledge correctness based on the calculated geographic knowledge graph to be measured crr 、Knowledge Error Indicator cnd Two indicators measure the content quality of the geographic knowledge graph to be measured con :

[0168]

[0169] Step 4: Calculate the quality of the application phase app ; Quality of the application stage app The method is to determine the degree to which the knowledge contained in the geographic knowledge graph to be determined meets the application requirements; specifically, the method comprises the following steps:

[0170] Step 4-1, calculating the conceptual integrity, spatial integrity, and temporal integrity of the knowledge in the geographic knowledge graph to be measured;

[0171] Step 4-2: Calculate the knowledge integrity indicator Indicator of the geographic knowledge graph to be measured cmp ; The knowledge integrity indicator Indicator cmp The quality of the application phase is determined by whether the geographic knowledge graph to be determined contains the target concept required by the application. The higher the knowledge integrity of the geographic knowledge graph to be determined, the higher the quality of the application phase. The knowledge integrity indicator Indicator cmp The calculation formula is:

[0172] Indicator cmp = cmp core ×cmp spatial ×cmp temporal

[0173] in,

[0174] cmp core Concept completeness, which measures the extent to which the geographic knowledge graph to be measured contains the target concept;

[0175] cmp spatial Spatial completeness, which measures the degree to which the geographic knowledge graph to be measured contains the spatial scope of the target concept;

[0176] cmp temporal Temporal completeness, which measures the extent to which the geographic knowledge graph to be measured contains the temporal range of the target concept;

[0177] Furthermore, for concept completeness, if the geographic knowledge graph to be measured contains the target concept, then cmp core The value is 1.00; if the geographic knowledge graph to be measured does not contain the target concept, cmp core The value is 0.50;

[0178] Furthermore, for spatial completeness, if the geographic knowledge graph to be measured contains the spatial scope of the target concept, then cmp spatial The value is 1.00; if the geographic knowledge graph to be measured does not contain the spatial scope of the target concept, cmp spatial The value is 0.50;

[0179] Furthermore, for temporal completeness, if the geographic knowledge graph to be measured contains the time range of the target concept, then cmp temporal The value is 1.00; the geographic knowledge graph to be measured does not contain the time range of the target concept, then cmp temporal The value is 0.50;

[0180] Furthermore, if the application target of the geographic knowledge graph to be measured is not specified, the relevant concepts, space, and time of the general knowledge graph are used for measurement;

[0181] Step 4-3: Calculate the concept representativeness rep of the knowledge in the geographic knowledge graph to be measured core , spatial representativeness rep spatial , time representativeness rep temporal ;

[0182] Furthermore, the concept representative rep core The following calculation formula can be used:

[0183]

[0184] in,

[0185] kg_num concept is the number of target concepts in the geographic knowledge graph to be measured;

[0186] KG_num concept is the number of target concepts in the reference domain knowledge graph;

[0187] The more target concepts in the reference domain knowledge graph contained in the geographic knowledge graph to be measured, the more representative the concept rep core The bigger;

[0188] Furthermore, the spatial representativeness rep spatial The following calculation formula can be used:

[0189]

[0190] in,

[0191] num sp_tar_trp is the amount of knowledge in the target geographic space area in the geographic knowledge graph to be measured;

[0192] num sp_gen_trp is the amount of knowledge in the target geographic space area in the reference domain knowledge graph;

[0193] The more target concepts in the reference domain knowledge graph contained in the target geographic spatial area, the more representative the spatial representation. spatial The bigger;

[0194] Furthermore, the time representativeness rep temporal The following calculation formula can be used:

[0195]

[0196] in,

[0197] num tp_tar_trp is the amount of knowledge in the target time range in the geographic knowledge graph to be measured;

[0198] num tp_gen_trp It is the amount of knowledge in the reference domain knowledge graph within the target time range;

[0199] The more target concepts in the reference domain knowledge graph contained in the geographic knowledge graph to be measured within the target time range, the more representative the time representation is. temporal The bigger;

[0200] Step 4-4: Calculate the knowledge representativeness indicator Indicator in the geographic knowledge graph to be measured rep ; The knowledge representativeness indicator rep The quality of the application phase is measured by using the distribution deviation between the knowledge contained in the geographic knowledge graph to be measured and the real-world knowledge. The higher the knowledge representativeness of the geographic knowledge graph to be measured, the higher the quality of the application phase. The knowledge representativeness indicator Indicator rep The calculation formula is

[0201] Indicator rep =rep core ×rep spatial ×rep temporal

[0202] in,

[0203] rep core The representativeness of the concept is to measure the degree to which the target concept contained in the geographic knowledge graph to be measured represents the real-world knowledge;

[0204] rep spatial Spatial representativeness, measuring the degree to which the target concept contained in the geographic knowledge graph to be measured represents the real-world knowledge in space;

[0205] rep temporalTemporal representativeness, measuring the temporal representativeness of the target concept contained in the geographic knowledge graph to be measured to the real-world knowledge;

[0206] Step 4-5: Use the Knowledge Completeness Indicator cmp and Knowledge Representativeness Indicator r□□ Two indicators measure the application quality of the geographic knowledge graph to be measured app :

[0207]

[0208] Step 5: Calculate the quality of the exchange phase acc ; Quality of the exchange phase acc The method is to determine the degree to which the geographic knowledge graph to be determined satisfies external data exchange; specifically, the following steps are included:

[0209] Step 5-1: Calculate the comprehensibility cph in the geographic knowledge graph to be measured understand , operability cph interoperate ;

[0210] Furthermore, the comprehensibility cph understand The following calculation formula can be used:

[0211] cph understand =schema×document

[0212] in,

[0213] schema is URI accessibility. If the URI used to access the geographic knowledge graph to be determined is valid, the schema value is 1.00; if the URI used to access the geographic knowledge graph to be determined is invalid, the schema value is 0.00;

[0214] document is the description document. If the geographic knowledge graph to be measured has a supporting description document, the value of document is 1.00; if the geographic knowledge graph to be measured does not have a supporting description document, the value of document is 0.00;

[0215] Furthermore, the interoperability cph interoperate The following calculation formula can be used:

[0216]

[0217] in,

[0218] num unknown_conceptis the number of concepts in the geographic knowledge graph to be determined that are unknown to the existing ontology;

[0219] hum concept is the number of all concepts in the geographic knowledge graph to be measured;

[0220] format is the case of using the knowledge graph standard format. If the geographic knowledge graph to be measured is exchanged using the knowledge graph standard format such as RDF, N-Triples, Turtle, JSON-LD, etc., format takes a value of 1.00; if the geographic knowledge graph to be measured is not exchanged using the knowledge graph standard format, format takes a value of 0.00;

[0221] Step 5-2: Calculate the interpretability indicator Indicator in the geographic knowledge graph to be measured cph ; The interpretability indicator Indicator cph The quality of the exchange phase is measured by the degree to which the geographic knowledge graph to be measured can be understood by humans and computers without obstacles. The higher the interpretability of the geographic knowledge graph to be measured, the higher the quality of the exchange phase. The interpretability indicator Indicator cph The calculation formula is:

[0222] Indicator cph =cph understand ×cph interoperate

[0223] in,

[0224] cph understand It is comprehensibility;

[0225] cph interoperate It is operability;

[0226] Step 5-3: Calculate the availability of the geographic knowledge graph to be determined exc available , interconnectivity exc interlink , permission level exc licensing ;

[0227] Furthermore, the availability exc available The following calculation formula can be used:

[0228] exc available =volume available / volume total

[0229] in,

[0230] volume availableis the amount of data in the geographic knowledge graph to be determined that can be downloaded by visitors;

[0231] volume total is the total amount of data in the geographic knowledge graph to be measured;

[0232] The more data in the geographic knowledge graph to be determined that can be downloaded by visitors, the greater the availability. available The bigger;

[0233] For the interconnectivity exc interlink The following calculation formula can be used:

[0234] exc interlink =num ins_linking / num ins

[0235] in,

[0236] num ins_linking is the number of instances with external links in the geographic knowledge graph to be determined;

[0237] num ins is the number of all instances in the geographic knowledge graph to be measured;

[0238] The more instances with external links in the geographic knowledge graph to be determined, the greater the interconnectivity interlink The bigger;

[0239] Furthermore, the permission level exc licensing If the geographic knowledge graph to be determined provides a data authorization use condition file, then exc licensing The value is 1.00; if the geographic knowledge graph to be measured does not provide a data authorization use condition file, then exc licensing The value is 0.50;

[0240] Step 5-4: Calculate the exchangeability indicator Indicator of the geographic knowledge graph to be measured exc ; The interchangeability indicator Indicator exc The quality of the exchange phase is measured by the degree to which the geographic knowledge graph to be measured can be accessed and obtained. The higher the exchangeability of the geographic knowledge graph to be measured, the higher the quality of the exchange phase. The exchangeability indicator Indicator exc The calculation formula is:

[0241] Indicator exc =exc available ×exc interlink ×exc licensing

[0242] in,

[0243] exc available It is availability;

[0244] exc interlink It is interconnectedness;

[0245] exc licensing is the level of permission;

[0246] Step 5-5. Use Interpretability Indicator cph and Interchangeability Indicator exc Two indicators measure the quality of the exchange phase of the geographic knowledge graph to be measured: acc :

[0247]

[0248] Step 6: Based on steps 1 to 5, calculate the quality of the geographic knowledge graph to be measured GeoKG , the formula is

[0249]

[0250] In the formula,

[0251] Quality ins For data stage quality;

[0252] Quality mth For the quality of production stage;

[0253] Quality con For content stage quality;

[0254] Quality app Quality in the application phase;

[0255] Quality acc For the exchange stage quality;

[0256] Using the above scheme, the method for determining the quality of geographic knowledge graph based on life cycle theory disclosed in the present invention has the following advantages:

[0257] The method for measuring the quality of a geographic knowledge graph based on the life cycle theory of the present invention divides the life cycle of a geographic knowledge graph into data stage, production stage, content stage, application stage and exchange stage. According to the definition of the geographic knowledge graph life cycle, the quality of the geographic knowledge graph is comprehensively measured by the quality of the data stage, the quality of the production stage, the quality of the content stage, the quality of the application stage and the quality of the exchange stage, thus forming a complete quality measurement framework that can correctly reflect the quality of the geographic knowledge graph.

[0258] The concept, specific technical scheme and technical effects of the present invention will be further described below in conjunction with specific implementation methods to fully understand the purpose, characteristics and effects of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0259] Figure 1 It is a flow chart of a method for determining the quality of a geographic knowledge graph based on the life cycle theory of the present invention; DETAILED DESCRIPTION

[0260] The following introduces several preferred embodiments of the present invention to make its technical content clearer and easier to understand. The present invention can be embodied in many different forms of embodiments, which are exemplary descriptions, and the protection scope of the present invention is not limited to the embodiments mentioned in the text.

[0261] Example 1: A method for measuring the quality of a geographic knowledge graph based on life cycle theory

[0262] Step 1: Calculate the data quality ins ; Quality of the data stage ins The quality of the input data of the geographic knowledge graph to be measured includes the following steps:

[0263] Step 1-1, calculate the reliability of the knowledge source; the knowledge source reliability indicator Indicator prv The quality of the data stage is determined by using the reliability of the knowledge source for constructing the geographic knowledge graph to be determined, and the higher the reliability of the knowledge source for constructing the geographic knowledge graph to be determined, the higher the quality of the data stage;

[0264] The calculation of the reliability of the knowledge source specifically includes the following steps:

[0265] Step 1-1-1: Calculate the hierarchical weight prv of a knowledge source weight , the information level of knowledge source prv information ;

[0266] The information level prv of the knowledge source information The calculation formula is:

[0267]

[0268] in,

[0269] provider level is the registration level or reputation level index of the knowledge source;

[0270] provider number is the contribution index of the knowledge source;

[0271] provider aistance is the spatial deviation distance index between the knowledge source and the provided knowledge;

[0272] The registration level or reputation level index of the knowledge source provider level Use the following calculation formula to get:

[0273]

[0274] in,

[0275] Level(provider) is the registration level or reputation level of the knowledge source;

[0276] MAX_LEVEL is the maximum registration level or reputation level in all knowledge sources.

[0277] The higher the registration level or reputation level of the knowledge source, the higher the registration level or reputation level index provider le□□l The bigger;

[0278] Furthermore, the contribution index of the knowledge source provider number Use the following calculation formula to get:

[0279]

[0280] in,

[0281] number(provider) is the amount of knowledge contributed by the knowledge source;

[0282] AVERAGE_NUMBER is the average amount of knowledge contributed by all knowledge sources;

[0283] The more knowledge the knowledge source contributes, the higher the contribution index of the knowledge sourceprovider number The bigger;

[0284] The spatial deviation distance index between the knowledge source and the provided knowledge provider distance Use the following calculation formula to get:

[0285]

[0286] distance(providercity-object) is the Euclidean distance between the knowledge source and the provided knowledge in geographic space;

[0287] The smaller the Euclidean distance between the knowledge source and the provided knowledge in geographic space, the higher the spatial deviation distance index provider distance The bigger;

[0288] The hierarchical weight prv of the knowledge source weight Determined by provider tier;

[0289] If the provider level is a world or national authoritative data source, the level weight is 1.00;

[0290] The provider level is a provincial or regional high-confidence data source with a level weight of 0.75;

[0291] The provider level is a general public data source, and the level weight is 0.50;

[0292] The provider level is the personal data source, and the level weight is 0.25;

[0293]

[0294] Step 1-1-2: Calculate the source reliability Prv of one piece of knowledge among the n pieces of knowledge in the geographic knowledge graph to be measured i ; The source reliability of the piece of knowledge Prv i The calculation formula is:

[0295] Prv i =prv weight ×prv information

[0296] in,

[0297] prv weight is the hierarchical weight of the knowledge source;

[0298] prv information is the information level of the knowledge source;

[0299] Step 1-1-3: Calculate the source reliability indicator prv , the knowledge source reliability indicator Indicator prv The calculation formula is:

[0300]

[0301] in,

[0302] The geographic knowledge graph to be determined is constructed from n pieces of knowledge;

[0303] Prv i The reliability of the source of one piece of knowledge;

[0304] Step 1-2: Calculate the knowledge source spatiotemporal coverage indicator crg ; The knowledge source spatiotemporal coverage indicator Indicator crg The data stage quality is measured using the spatial and temporal restriction degree of the knowledge source for constructing the geographic knowledge graph to be measured. The higher the spatiotemporal coverage of the knowledge source for constructing the geographic knowledge graph to be measured, the higher the data stage quality; specifically comprising the following steps:

[0305] Step 1-2-1: Calculate the spatial coverage crg of one piece of knowledge among the n pieces of knowledge constituting the geographic knowledge graph to be measured spatial , time coverage crg temporal ;

[0306] The spatial coverage crg spatial Use the following calculation formula to get:

[0307]

[0308] in,

[0309] num spatial is the number of spatial knowledge triples in the geographic knowledge graph to be determined;

[0310] num sp_limited is the number of spatial knowledge triples with spatial area restrictions in the geographic knowledge graph to be determined;

[0311] The fewer spatial area restrictions there are in the geographic knowledge graph to be measured, the greater the spatial coverage crg spatial The bigger;

[0312] The time coverage crg temporal Use the following calculation formula to get:

[0313]

[0314] in,

[0315] num temporal is the number of knowledge triples in the geographic knowledge graph to be determined;

[0316] num tp_limitedis the number of spatial knowledge triples with time range restrictions in the geographic knowledge graph to be determined;

[0317] The fewer time range restrictions there are in the geographic knowledge graph to be measured, the greater the time coverage crg temporal The bigger;

[0318] Step 1-2-2: Calculate the spatiotemporal coverage of a piece of knowledge source Crg i ; The calculation formula is:

[0319] Cr i =crg spatial ×crg temporal

[0320] in,

[0321] crg spatial is the spatial coverage, which measures the coverage of the knowledge source in space;

[0322] crg temporal is the temporal coverage, which measures the temporal coverage of the knowledge source;

[0323] Step 1-2-3: Calculate the knowledge source spatiotemporal coverage indicator crg ; The calculation formula is:

[0324]

[0325] in,

[0326] Cr i is the spatiotemporal coverage of the knowledge source of one piece of knowledge among the n pieces of knowledge constituting the geographic knowledge graph to be measured;

[0327] Step 1-3: Calculate the knowledge medium reliability indicator med ; The knowledge medium reliability indicator Indicator med The data stage quality is measured using the reliability of the transmission medium that carries the knowledge for constructing the geographic knowledge graph to be measured. The higher the reliability of the knowledge medium for constructing the geographic knowledge graph to be measured, the higher the data stage quality. Specifically, the following steps are included:

[0328] Step 1-3-1, calculate the knowledge medium suitability med of a piece of knowledge in the said generation determination geographic knowledge graph suitability , media loss type ;

[0329] The knowledge medium suitability suitability Use the following calculation formula to get:

[0330]

[0331] in,

[0332] The geographic knowledge graph to be determined is constructed from l types of knowledge;

[0333] st i the suitability of the type of medium for one of the categories of knowledge;

[0334] If such knowledge is transmitted through a suitable medium, then the st i The value is 1.00;

[0335] If such knowledge is transmitted through an inappropriate medium, the st i The value is 0.50;

[0336] Furthermore, the media loss med type , determined according to the type of knowledge;

[0337] The knowledge type is structured knowledge, the medium loss is lossless, and the medium lossiness is 1.00;

[0338] The knowledge type is semi-structured knowledge, the media loss is lossy, and the media lossiness is 0.75;

[0339] The knowledge type is unstructured knowledge, the media loss is ambiguous, and the media loss is 0.50;

[0340] Type of knowledge Media loss <![CDATA[med type ]]> Structured Knowledge Lossless 1.00 Semi-structured knowledge Lossy 0.75 Unstructured knowledge Create ambiguity 0.50

[0341] Step 1-3-2: Calculate the media reliability Med of a piece of knowledge i The formula is

[0342] Med i =med suitability ×med type

[0343] in,

[0344] med suitability is the suitability of the medium of knowledge, which determines whether the knowledge is transmitted through the most appropriate medium;

[0345] med type The medium lossiness is to determine whether the medium used by the knowledge causes transmission loss to the knowledge;

[0346] Step 1-3-3: Calculate the knowledge medium reliability indicator med ; The knowledge medium reliability indicator Indicator med The calculation formula is:

[0347]

[0348] in,

[0349] Med i is the medium reliability of one piece of knowledge among the n pieces of knowledge constituting the geographic knowledge graph to be measured;

[0350] Step 1-4: Knowledge source reliability indicator obtained based on steps 1-1 to 1-3 prv 、Knowledge source spatiotemporal coverage indicator crg 、Knowledge medium reliability indicator med The three indicators comprehensively measure the quality of the data to be measured. ins The formula is

[0351]

[0352] In the formula,

[0353] Indicator prv Reliability of knowledge sources;

[0354] Indicator crg is the spatial and temporal coverage of knowledge sources;

[0355] Indicator med Reliability as a medium of knowledge;

[0356] Step 2: Calculate the quality of the production stage mth ; Quality of the production stage mth The method is to determine the quality of the input data when it is converted into the geographic knowledge graph to be determined, which specifically includes the following steps:

[0357] Step 2-1: Determine the performance of the conversion method performance , conversion method confidence mth type ;

[0358] The conversion method performance perfomance The F1 value in the field of natural language processing can be used for calculation:

[0359]

[0360] in,

[0361] P is the accuracy of the conversion result: P = number of correctly converted data / total number of converted data;

[0362] R is the recall of the conversion result: R = number of correctly converted data / number of data that should be converted;

[0363] The conversion method confidence mth type This can be determined using the test level of the conversion result:

[0364] mth type =1-∝

[0365] in,

[0366] ∝ is the test level or significance level of the result;

[0367] Step 2-2: Calculate the method reliability Mth of a conversion method i ; The calculation formula is:

[0368] M i =mth performance ×mth type

[0369] in,

[0370] mth performance is the conversion method performance;

[0371] mth type is the conversion method confidence;

[0372] Step 2-3: Determine the production quality of the geographic knowledge graph to be determined mth , the calculation formula is:

[0373]

[0374] in,

[0375] M i reliability of a method for converting the input data into one of the conversion methods for the geographic knowledge graph to be determined using m conversion methods;

[0376] Step 3: Calculate the quality of the content stage con ; Quality of the content stage con The method is to determine the quality of the knowledge contained in the geographic knowledge graph to be determined; specifically, the method comprises the following steps:

[0377] Step 3-1: Calculate the knowledge logic correctness crr of a piece of knowledge in the geographic knowledge graph to be measured logic , knowledge content correctness crr content ;

[0378] For the logical correctness of the knowledge crrlogic The following calculation formula can be used:

[0379]

[0380] in,

[0381] num wl is the number of knowledge in the geographic knowledge graph to be determined that conflicts with the ontology model;

[0382] n is the number of all knowledge in the geographic knowledge graph to be measured;

[0383] The less the knowledge in the geographic knowledge graph to be measured that conflicts with the ontology model, the more correct the knowledge logic is. logic The bigger;

[0384] The correctness of the knowledge content content The following calculation formula can be used:

[0385]

[0386] in,

[0387] num wt is the amount of knowledge in the geographic knowledge graph to be measured that is inconsistent with the actual situation;

[0388] n is the number of all knowledge in the geographic knowledge graph to be measured;

[0389] The less the knowledge in the geographic knowledge graph to be measured is inconsistent with the actual situation, the more correct the knowledge content is. content The bigger;

[0390] Step 3-2: Calculate the knowledge correctness Crr of a piece of knowledge i ; The calculation formula is:

[0391] Crr i =crr logic ×crr content

[0392] in,

[0393] crr loqic It is the logical correctness of knowledge;

[0394] crr content It is the correctness of knowledge content;

[0395] Step 3-3: Calculate the knowledge correctness indicator of the geographic knowledge graph to be measured crr ; The knowledge correctness indicator Indicator crrThe correctness of the knowledge contained in the geographic knowledge graph to be measured is used to measure the quality of the content stage. The higher the correctness of the knowledge in the geographic knowledge graph to be measured, the higher the quality of the content stage. The knowledge correctness indicator Indicator crr The calculation formula is

[0396]

[0397] in,

[0398] Crr i is the knowledge correctness of one piece of knowledge among the n pieces of knowledge constituting the geographic knowledge graph to be measured;

[0399] Step 3-4: Calculate the category error cnd of the geographic knowledge graph knowledge to be measured type , spatial error of knowledge cnd spatial , the time error of knowledge cnd temporal ;

[0400] The category error cnd type The following calculation formula can be used:

[0401]

[0402] in,

[0403] num wtt is the number of knowledge with incorrect categories in the geographic knowledge graph to be measured;

[0404] n is the number of all knowledge in the geographic knowledge graph to be measured;

[0405] The smaller the number of knowledge with category errors in the geographic knowledge graph to be measured, the smaller the category error cnd type The bigger;

[0406] The spatial error cnd spatial The following calculation formula can be used:

[0407]

[0408] in,

[0409] err_spatial i is the number of erroneous knowledge in the geographic knowledge graph to be measured at a certain spatial granularity;

[0410] n is the number of all knowledge in the geographic knowledge graph to be measured, and the spatial granularity and its number p can be set with reference to different spatial division methods such as administrative divisions, spatial resolution, and scale;

[0411] The smaller the number of erroneous knowledge in the geographic knowledge graph to be measured in different spatial granularities, the smaller the spatial error cnd spatial The bigger;

[0412] The time error cnd temporal The following calculation formula can be used:

[0413]

[0414] in,

[0415] err_temporal i is the number of erroneous knowledge in the geographic knowledge graph to be measured at a certain time granularity;

[0416] n is the number of all knowledge in the geographic knowledge graph to be measured, and the time granularity and its number q can be set in years, months, days, hours, minutes, seconds, etc.;

[0417] The smaller the number of erroneous knowledge in the geographic knowledge graph to be measured in different time granularities, the smaller the time error cnd teraloral The bigger;

[0418] Step 3-5: Calculate the knowledge error indicator Indicator of the geographic knowledge graph knowledge to be measured cnd ; The knowledge error indicator Indicator cnd The content stage quality is measured using the error degree of the knowledge contained in the geographic knowledge graph to be measured. The lower the knowledge error of the geographic knowledge graph to be measured, the higher the content stage quality. The knowledge error indicator Indicator cnd The calculation formula is:

[0419] Indicator cnd =cnd type ×cnd spatial ×cnd temporal

[0420] in,

[0421] cnd type is the category error of the n pieces of knowledge constituting the geographic knowledge graph to be measured;

[0422] cnd spatial is the spatial error of the n pieces of knowledge constituting the geographic knowledge graph to be measured;

[0423] cnd temporal is the time error of the n pieces of knowledge constituting the geographic knowledge graph to be measured;

[0424] Step 3-6: Indicator of knowledge correctness based on the calculated geographic knowledge graph to be measured crr 、Knowledge Error Indicator cnd Two indicators measure the content quality of the geographic knowledge graph to be measured con :

[0425]

[0426] Step 4: Calculate the quality of the application phase app ; Quality of the application stage app The method is to determine the degree to which the knowledge contained in the geographic knowledge graph to be determined meets the application requirements; specifically, the method comprises the following steps:

[0427] Step 4-1, calculating the conceptual integrity, spatial integrity, and temporal integrity of the knowledge in the geographic knowledge graph to be measured;

[0428] Step 4-2: Calculate the knowledge integrity indicator Indicator of the geographic knowledge graph to be measured cmp ; The knowledge integrity indicator Indicator cmp The quality of the application phase is determined by whether the geographic knowledge graph to be determined contains the target concept required by the application. The higher the knowledge integrity of the geographic knowledge graph to be determined, the higher the quality of the application phase. The knowledge integrity indicator Indicator cmp The calculation formula is:

[0429] Indicator cmp = cmp core ×cmp spatial ×cmp temporal

[0430] in,

[0431] cmp core Concept completeness, which measures the extent to which the geographic knowledge graph to be measured contains the target concept;

[0432] cmp spatial Spatial completeness, which measures the degree to which the geographic knowledge graph to be measured contains the spatial scope of the target concept;

[0433] cmp temporal Temporal completeness, which measures the extent to which the geographic knowledge graph to be measured contains the temporal range of the target concept;

[0434] For concept completeness, if the geographic knowledge graph to be measured contains the target concept, then cmp coreThe value is 1.00; if the geographic knowledge graph to be measured does not contain the target concept, cmp core The value is 0.50;

[0435] For spatial completeness, if the geographic knowledge graph to be measured contains the spatial scope of the target concept, then cmp spatial The value is 1.00; if the geographic knowledge graph to be measured does not contain the spatial scope of the target concept, cmp spatial The value is 0.50;

[0436] For temporal completeness, if the geographic knowledge graph to be measured contains the time range of the target concept, then cmp temporal The value is 1.00; the geographic knowledge graph to be measured does not contain the time range of the target concept, then cmp temporal The value is 0.50;

[0437] If the application target of the geographic knowledge graph to be measured is not specified, the relevant concepts, space, and time of the general knowledge graph are used for measurement;

[0438] Step 4-3: Calculate the concept representativeness rep of the knowledge in the geographic knowledge graph to be measured core , spatial representativeness rep spatial , time representativeness rep temporal ;

[0439] The concept is representative core The following calculation formula can be used:

[0440]

[0441] in,

[0442] kg_num concept is the number of target concepts in the geographic knowledge graph to be measured;

[0443] KG_num concept is the number of target concepts in the reference domain knowledge graph;

[0444] The more target concepts in the reference domain knowledge graph contained in the geographic knowledge graph to be measured, the more representative the concept rep core The bigger;

[0445] The spatial representation rep spatial The following calculation formula can be used:

[0446]

[0447] in,

[0448] num sp_ter_trp is the amount of knowledge in the target geographic space area in the geographic knowledge graph to be measured;

[0449] num sp_gen_trp is the amount of knowledge in the target geographic space area in the reference domain knowledge graph;

[0450] The more target concepts in the reference domain knowledge graph contained in the target geographic spatial area, the more representative the spatial representation. spatial The bigger;

[0451] The time representative rep temporal The following calculation formula can be used:

[0452]

[0453] in,

[0454] num tp_tartrp is the amount of knowledge in the target time range in the geographic knowledge graph to be measured;

[0455] num tp_gen_trp It is the amount of knowledge in the reference domain knowledge graph within the target time range;

[0456] The more target concepts in the reference domain knowledge graph contained in the geographic knowledge graph to be measured within the target time range, the more representative the time representation is. temporal The bigger;

[0457] Step 4-4: Calculate the knowledge representativeness indicator Indicator in the geographic knowledge graph to be measured rep ; The knowledge representativeness indicator rep The quality of the application phase is measured by using the distribution deviation between the knowledge contained in the geographic knowledge graph to be measured and the real-world knowledge. The higher the knowledge representativeness of the geographic knowledge graph to be measured, the higher the quality of the application phase. The knowledge representativeness indicator Indicator rep The calculation formula is

[0458] Indicator rep =rep core ×rep spatial ×rep temporal

[0459] in,

[0460] rep coreThe representativeness of the concept is to measure the degree to which the target concept contained in the geographic knowledge graph to be measured represents the real-world knowledge;

[0461] rep spatial Spatial representativeness, measuring the degree to which the target concept contained in the geographic knowledge graph to be measured represents the real-world knowledge in space;

[0462] rep temporal Temporal representativeness, measuring the temporal representativeness of the target concept contained in the geographic knowledge graph to be measured to the real-world knowledge;

[0463] Step 4-5: Use the Knowledge Completeness Indicator cmp and Knowledge Representativeness Indicator rep Two indicators measure the application quality of the geographic knowledge graph to be measured app :

[0464]

[0465] Step 5: Calculate the quality of the exchange phase acc ; Quality of the exchange phase acc The method is to determine the degree to which the geographic knowledge graph to be determined satisfies external data exchange; specifically, the following steps are included:

[0466] Step 5-1: Calculate the comprehensibility cph in the geographic knowledge graph to be measured understand , operability cph interoperate ;

[0467] The comprehensibility cph understand The following calculation formula can be used:

[0468] cph understand =schema:×document

[0469] in,

[0470] schema is URI accessibility. If the URI used to access the geographic knowledge graph to be determined is valid, the schema value is 1.00; if the URI used to access the geographic knowledge graph to be determined is invalid, the schema value is 0.00;

[0471] document is the description document. If the geographic knowledge graph to be measured has a supporting description document, the value of document is 1.00; if the geographic knowledge graph to be measured does not have a supporting description document, the value of document is 0.00;

[0472] The interoperability cph interoperate The following calculation formula can be used:

[0473]

[0474] in,

[0475] num unknown_concept is the number of concepts in the geographic knowledge graph to be determined that are unknown to the existing ontology;

[0476] hum concept is the number of all concepts in the geographic knowledge graph to be measured;

[0477] format is the case of using the knowledge graph standard format. If the geographic knowledge graph to be measured is exchanged using the knowledge graph standard format such as RDF, N-Triples, Turtle, JSON-LD, etc., format takes a value of 1.00; if the geographic knowledge graph to be measured is not exchanged using the knowledge graph standard format, format takes a value of 0.00;

[0478] Step 5-2: Calculate the interpretability indicator Indicator in the geographic knowledge graph to be measured cph ; The interpretability indicator Indicator cph The quality of the exchange phase is measured by the degree to which the geographic knowledge graph to be measured can be understood by humans and computers without obstacles. The higher the interpretability of the geographic knowledge graph to be measured, the higher the quality of the exchange phase. The interpretability indicator Indicator cph The calculation formula is:

[0479] Indicator cph =cph understand ×cph interoperate

[0480] in,

[0481] cph understand It is comprehensibility;

[0482] cph interoperate It is operability;

[0483] Step 5-3: Calculate the availability of the geographic knowledge graph to be determined exc available , interconnectivity exc interlink , permission level exc licensing ;

[0484] The availability exc available The following calculation formula can be used:

[0485] exc available =volume available / volume total

[0486] in,

[0487] volume available is the amount of data in the geographic knowledge graph to be determined that can be downloaded by visitors;

[0488] volume toral is the total amount of data in the geographic knowledge graph to be measured;

[0489] The more data in the geographic knowledge graph to be determined that can be downloaded by visitors, the greater the availability. auailable The bigger;

[0490] For the interconnectivity exc interlink The following calculation formula can be used:

[0491] exc interlink =num ins_linking / num ins

[0492] in,

[0493] num ins_linking is the number of instances with external links in the geographic knowledge graph to be determined;

[0494] num ins is the number of all instances in the geographic knowledge graph to be measured;

[0495] The more instances with external links in the geographic knowledge graph to be determined, the greater the interconnectivity interlink The bigger;

[0496] The degree of permission exc licensing If the geographic knowledge graph to be determined provides a data authorization use condition file, then exc licensing The value is 1.00; if the geographic knowledge graph to be measured does not provide a data authorization use condition file, then exc licensing The value is 0.50;

[0497] Step 5-4: Calculate the exchangeability indicator Indicator of the geographic knowledge graph to be measured exc ; The interchangeability indicator Indicator excThe quality of the exchange phase is measured by the degree to which the geographic knowledge graph to be measured can be accessed and obtained. The higher the exchangeability of the geographic knowledge graph to be measured, the higher the quality of the exchange phase. The exchangeability indicator Indicator exc The calculation formula is:

[0498] Indicator exc =exc available :×exc interlink :×exc licensing

[0499] in,

[0500] exc available It is availability;

[0501] exc interlink It is interconnectedness;

[0502] exc licensing is the level of permission;

[0503] Step 5-5. Use Interpretability Indicator cph and Interchangeability Indicator exc Two indicators measure the quality of the exchange phase of the geographic knowledge graph to be measured: acc :

[0504]

[0505] Step 6: Based on steps 1 to 5, calculate the quality of the geographic knowledge graph to be measured GeoKG , the formula is

[0506]

[0507] In the formula,

[0508] Quality ins For data stage quality;

[0509] Quality mth For the quality of production stage;

[0510] Quality con For content stage quality;

[0511] Quality app Quality in the application phase;

[0512] Quality acc For the exchange stage quality;

[0513] To summarize, the technical solution of this patent divides the life cycle of the geographic knowledge graph into data stage, production stage, content stage, application stage and exchange stage. According to the definition of the geographic knowledge graph life cycle, the quality of the geographic knowledge graph is comprehensively measured by the data stage quality, production stage quality, content stage quality, application stage quality and exchange stage quality, forming a complete quality measurement framework that can correctly reflect the quality of the geographic knowledge graph.

[0514] The preferred specific embodiments of the present invention are described in detail above. It should be understood that ordinary technicians in the field can make many modifications and changes based on the concept of the present invention without creative work. Therefore, all technical solutions that can be obtained by technicians in the technical field based on the concept of the present invention through logical analysis, reasoning or limited experiments on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. A method for determining the quality of a geographic knowledge graph based on life cycle theory, characterized in that: It includes dividing the life cycle of geographic knowledge graph into data stage, production stage, content stage, application stage and exchange stage, and defining the quality of geographic knowledge graph according to the life cycle of geographic knowledge graph. GeoKG , the quality of the geographic knowledge graph GeoKG Data stage quality ins , Quality in the production stage mth , Content stage quality con , Application stage quality app and exchange phase quality acc Comprehensive determination; specifically includes the following steps, Step 1: Calculate the data quality ins ; Quality of the data stage ins The quality of the input data of the geographic knowledge graph to be measured; Step 2: Calculate the quality of the production stage mth ; Quality of the production stage mth is to measure the quality of the input data when it is converted into the geographic knowledge graph to be measured; Step 3: Calculate the quality of the content stage con ; Quality of the content stage con Determine the quality of the knowledge contained in the geographic knowledge graph to be determined; Step 4: Calculate the quality of the application phase app ; Quality of the application stage app It is to determine the degree to which the knowledge contained in the geographic knowledge graph to be determined meets the application requirements; Step 5: Calculate the quality of the exchange phase acc ; Quality of the exchange phase acc It is to determine the degree to which the geographic knowledge graph to be determined satisfies external data exchange; Step 6: Based on steps 1 to 5, calculate the quality of the geographic knowledge graph to be measured; the formula is In the formula, Quality ins For data stage quality; Quality mth Quality at the production stage; Quality con For content stage quality; Quality app Quality in the application phase; Quality acc For the exchange stage quality; The step 1 specifically includes the following steps: Step 1-1: Calculate the knowledge source reliability indicator prv ; The knowledge source reliability indicator Indicator prv The quality of the data stage is determined by using the reliability of the knowledge source for constructing the geographic knowledge graph to be determined, and the higher the reliability of the knowledge source for constructing the geographic knowledge graph to be determined, the higher the quality of the data stage; Step 1-2: Calculate the knowledge source spatiotemporal coverage indicator crg ; The knowledge source spatiotemporal coverage indicator Indicator crg The data stage quality is measured by using the spatial and temporal restriction degree of the knowledge source for constructing the geographic knowledge graph to be measured, and the higher the spatiotemporal coverage of the knowledge source for constructing the geographic knowledge graph to be measured, the higher the data stage quality; Step 1-3: Calculate the knowledge medium reliability indicator med ; The knowledge medium reliability indicator Indicator med The data stage quality is measured using the reliability of the transmission medium that carries the knowledge for constructing the geographic knowledge graph to be measured, and the higher the reliability of the knowledge medium for constructing the geographic knowledge graph to be measured, the higher the data stage quality; Step 1-4: Knowledge source reliability indicator obtained based on steps 1-1 to 1-3 prv 、Knowledge source spatiotemporal coverage indicator crg 、Knowledge medium reliability indicator med Three indicators comprehensively measure the quality of the data stage ins The formula is In the formula, Indicator prv Reliability of knowledge sources; Indicator crg is the spatial and temporal coverage of knowledge sources; Indicator med Reliability as a medium of knowledge; The step 2 specifically includes the following steps: Step 2-1: Determine the performance of the conversion method performance , conversion method confidence mth type ; The conversion method performance performance Using the F1 value calculation in the field of natural language processing: in, P is the accuracy of the conversion result: P = number of correctly converted data / total number of converted data; R is the recall of the conversion result: R = number of correctly converted data / number of data that should be converted; The conversion method confidence mth type The test level of the transformation result is used for determination: mth type =1-∝ in, ∝ is the test level or significance level of the result; Step 2-2: Calculate the method reliability Mth of a conversion method i ; The calculation formula is: Mth i =mth performance ×mth type in, mth performance is the conversion method performance; mth type is the conversion method confidence; Step 2-3: Determine the production quality of the geographic knowledge graph to be determined mth , the calculation formula is: in, M i reliability of a method for converting the input data into one of the conversion methods for the geographic knowledge graph to be determined using m conversion methods; The step 3 specifically includes the following steps: Step 3-1: Calculate the knowledge logic correctness crr of a piece of knowledge in the geographic knowledge graph to be measured logic 、Correctness of knowledge content content ; For the logical correctness of the knowledge crr logic Use the following calculation formula to get: in, num wl is the number of knowledge in the geographic knowledge graph to be determined that conflicts with the ontology model; n is the number of all knowledge in the geographic knowledge graph to be measured; The less the knowledge in the geographic knowledge graph to be measured that conflicts with the ontology model, the more correct the knowledge logic is. logic The bigger; The correctness of the knowledge content content Use the following calculation formula to get: in, num wt is the amount of knowledge in the geographic knowledge graph to be measured that is inconsistent with the actual situation; n is the number of all knowledge in the geographic knowledge graph to be measured; The less the knowledge in the geographic knowledge graph to be measured is inconsistent with the actual situation, the more correct the knowledge content is. content The bigger; Step 3-2: Calculate the knowledge correctness Crr of a piece of knowledge i ; The calculation formula is: Crr i =crr logic ×crr content in, crr logic It is the logical correctness of knowledge; crr content It is the correctness of knowledge content; Step 3-3: Calculate the knowledge correctness indicator of the geographic knowledge graph to be measured crr ; The knowledge correctness indicator Indicator crr The correctness of the knowledge contained in the geographic knowledge graph to be measured is used to measure the quality of the content stage. The higher the correctness of the knowledge in the geographic knowledge graph to be measured, the higher the quality of the content stage. The knowledge correctness indicator Indicator crr The calculation formula is in, Crr i is the knowledge correctness of one piece of knowledge among the n pieces of knowledge constituting the geographic knowledge graph to be measured; Step 3-4: Calculate the category error cnd of the geographic knowledge graph knowledge to be measured type , spatial error of knowledge cnd spatial , the time error of knowledge cnd temporal ; The category error cnd type Use the following calculation formula to get: in, num wtt is the number of knowledge with incorrect categories in the geographic knowledge graph to be measured; n is the number of all knowledge in the geographic knowledge graph to be measured; The smaller the number of knowledge with category errors in the geographic knowledge graph to be measured, the smaller the category error cnd type The bigger; The spatial error cnd spatial Use the following calculation formula to get: in, err_spatial i is the number of erroneous knowledge in the geographic knowledge graph to be measured at a certain spatial granularity; n is the number of all knowledge in the geographic knowledge graph to be measured, and the spatial granularity and its number p are set with reference to different spatial division methods of administrative divisions, spatial resolution, and scale; The smaller the number of erroneous knowledge in the geographic knowledge graph to be measured in different spatial granularities, the smaller the spatial error cnd spatial The bigger; The time error cnd temporal Use the following calculation formula to get: in, err_temporal i is the number of erroneous knowledge in the geographic knowledge graph to be measured at a certain time granularity; n is the number of all knowledge in the geographic knowledge graph to be measured, and the time granularity and its number q are set in years, months, days, hours, minutes, and seconds; The smaller the number of erroneous knowledge in the geographic knowledge graph to be measured in different time granularities, the smaller the time error cnd temporal The bigger; Step 3-5: Calculate the knowledge error indicator Indicator of the geographic knowledge graph knowledge to be measured cnd ; The knowledge error indicator Indicator cnd The content stage quality is measured using the error degree of the knowledge contained in the geographic knowledge graph to be measured. The lower the knowledge error of the geographic knowledge graph to be measured, the higher the content stage quality. The knowledge error indicator Indicator cnd The calculation formula is: Indicator cnd =cnd type ×cnd spatial ×cnd temporal in, cnd type is the category error of the n pieces of knowledge constituting the geographic knowledge graph to be measured; cnd spatial is the spatial error of the n pieces of knowledge constituting the geographic knowledge graph to be measured; cnd temporal is the time error of the n pieces of knowledge constituting the geographic knowledge graph to be measured; Step 3-6: Indicator of knowledge correctness based on the calculated geographic knowledge graph to be measured crr 、Knowledge Error Indicator cnd Two indicators measure the content quality of the geographic knowledge graph to be measured con : The step 4 specifically includes the following steps: Step 4-1, calculating the conceptual integrity, spatial integrity, and temporal integrity of the knowledge in the geographic knowledge graph to be measured; Step 4-2: Calculate the knowledge integrity indicator Indicator of the geographic knowledge graph to be measured cmp ; The knowledge integrity indicator Indicator cmp The quality of the application phase is determined by whether the geographic knowledge graph to be determined contains the target concept required by the application. The higher the knowledge integrity of the geographic knowledge graph to be determined, the higher the quality of the application phase. The knowledge integrity indicator Indicator cmp The calculation formula is: Indicator cmp =cmp core ×cmp spatial ×cmp temporal in, cmp core Concept completeness, which measures the extent to which the geographic knowledge graph to be measured contains the target concept; cmp spatial Spatial completeness, which measures the degree to which the geographic knowledge graph to be measured contains the spatial scope of the target concept; cmp temporal Temporal completeness, which measures the extent to which the geographic knowledge graph to be measured contains the temporal range of the target concept; For concept completeness, if the geographic knowledge graph to be measured contains the target concept, then cmp core The value is 1.00; if the geographic knowledge graph to be measured does not contain the target concept, cmp core The value is 0.50; For spatial completeness, if the geographic knowledge graph to be measured contains the spatial scope of the target concept, then cmp spatial The value is 1.00; if the geographic knowledge graph to be measured does not contain the spatial scope of the target concept, cmp spatial The value is 0.50; For temporal completeness, if the geographic knowledge graph to be measured contains the time range of the target concept, then cmp temporal The value is 1.00; the geographic knowledge graph to be measured does not contain the time range of the target concept, then cmp temporal The value is 0.50; If the application target of the geographic knowledge graph to be measured is not specified, the relevant concepts, space, and time of the general knowledge graph are used for measurement; Step 4-3: Calculate the concept representativeness rep of the knowledge in the geographic knowledge graph to be measured core , spatial representativeness rep spatial , time representativeness rep temporal ; The concept is representative core Use the following calculation formula to get: in, kg_num concept is the number of target concepts in the geographic knowledge graph to be measured; KG_num concept is the number of target concepts in the reference domain knowledge graph; The more target concepts in the reference domain knowledge graph contained in the geographic knowledge graph to be measured, the more representative the concept rep core The bigger; The spatial representation rep spatial Use the following calculation formula to get: in, num sp_tar_trp is the amount of knowledge in the target geographic space area in the geographic knowledge graph to be measured; num sp_gen_trp is the amount of knowledge in the target geographic space area in the reference domain knowledge graph; The more target concepts in the reference domain knowledge graph contained in the target geographic spatial area, the more representative the spatial representation. spatial The bigger; The time representative rep temporal Use the following calculation formula to get: in, num tp_tar_trp is the amount of knowledge in the target time range in the geographic knowledge graph to be measured; num tp_gen_trp It is the amount of knowledge in the reference domain knowledge graph within the target time range; The more target concepts in the reference domain knowledge graph contained in the geographic knowledge graph to be measured within the target time range, the more representative the time representation is. temporal The bigger; Step 4-4: Calculate the knowledge representativeness indicator Indicator in the geographic knowledge graph to be measured rep ; The knowledge representativeness indicator rep The quality of the application phase is measured by using the distribution deviation between the knowledge contained in the geographic knowledge graph to be measured and the real-world knowledge. The higher the knowledge representativeness of the geographic knowledge graph to be measured, the higher the quality of the application phase. The knowledge representativeness indicator Indicator rep The calculation formula is Indicator rep =rep core ×rep spatial ×rep temporal in, rep core The representativeness of the concept is to measure the degree to which the target concept contained in the geographic knowledge graph to be measured represents the real-world knowledge; rep spatial Spatial representativeness, measuring the degree to which the target concept contained in the geographic knowledge graph to be measured represents the real-world knowledge in space; rep temporal Temporal representativeness, measuring the temporal representativeness of the target concept contained in the geographic knowledge graph to be measured to the real-world knowledge; Step 4-5: Use the Knowledge Completeness Indicator cmp and Knowledge Representativeness Indicator rep Two indicators measure the application quality of the geographic knowledge graph to be measured app : The step 5 specifically includes the following steps: Step 5-1: Calculate the comprehensibility cph in the geographic knowledge graph to be measured understand , operability cph interoperate ; The comprehensibility cph understand Use the following calculation formula to get: cph understand =schema×document in, schema is URI accessibility. If the URI used to access the geographic knowledge graph to be determined is valid, the schema value is 1.00; if the URI used to access the geographic knowledge graph to be determined is invalid, the schema value is 0.00; document is the description document. If the geographic knowledge graph to be measured has a supporting description document, the value of document is 1.00; if the geographic knowledge graph to be measured does not have a supporting description document, the value of document is 0.00; The operability cph interoperate Use the following calculation formula to get: in, num unknown_concept is the number of concepts in the geographic knowledge graph to be determined that are unknown to the existing ontology; num concept is the number of all concepts in the geographic knowledge graph to be measured; format is the case of using the knowledge graph standard format. If the geographic knowledge graph to be measured is exchanged using the RDF, N-Triples, Turtle, or JSON-LD knowledge graph standard format, format takes a value of 1.00; if the geographic knowledge graph to be measured is not exchanged using the knowledge graph standard format, format takes a value of 0.00; Step 5-2: Calculate the interpretability indicator Indicator in the geographic knowledge graph to be measured cph ; The interpretability indicator Indicator cph The quality of the exchange phase is measured by the degree to which the geographic knowledge graph to be measured can be understood by humans and computers without obstacles. The higher the interpretability of the geographic knowledge graph to be measured, the higher the quality of the exchange phase. The interpretability indicator Indicator cph The calculation formula is: Indicator cph =cph understand ×cph interoperate in, cph understand It is comprehensibility; cph interoperate It is operability; Step 5-3: Calculate the availability of the geographic knowledge graph to be determined exc available , interconnectivity exc interlink , permission level exc licensing ; The availability exc available Use the following calculation formula to get: exc available =volume available / volume total in, volume available is the amount of data in the geographic knowledge graph to be determined that can be downloaded by visitors; volume total is the total amount of data in the geographic knowledge graph to be measured; The more data in the geographic knowledge graph to be determined that can be downloaded by visitors, the greater the availability. available The bigger; For the interconnectivity exc interlink Use the following calculation formula to get: except interlink =number ins_linking / number ins in, num ins_linking is the number of instances with external links in the geographic knowledge graph to be determined; num ins is the number of all instances in the geographic knowledge graph to be measured; The more instances with external links in the geographic knowledge graph to be determined, the greater the interconnectivity interlink The bigger; The degree of permission exc licensing If the geographic knowledge graph to be determined provides a data authorization use condition file, then exc licensing The value is 1.00; if the geographic knowledge graph to be measured does not provide a data authorization use condition file, then exc licensing The value is 0.50; Step 5-4: Calculate the exchangeability indicator Indicator of the geographic knowledge graph to be measured exc ; The interchangeability indicator Indicator exc The quality of the exchange phase is measured by the degree to which the geographic knowledge graph to be measured can be accessed and obtained. The higher the exchangeability of the geographic knowledge graph to be measured, the higher the quality of the exchange phase. The exchangeability indicator Indicator exc The calculation formula is: Indicator exc =exc available ×exc interlink ×exc licensing in, exc available It is availability; exc interlink It is interconnectedness; exc licensing is the level of permission; Step 5-5. Use Interpretability Indicator cph and Interchangeability Indicator exc Two indicators measure the quality of the exchange phase of the geographic knowledge graph to be measured: acc :

2. The method for determining the quality of a geographic knowledge graph based on the life cycle theory according to claim 1, wherein in step 1-1, The calculation of the reliability of the knowledge source specifically includes the following steps: Step 1-1-1: Calculate the hierarchical weight prv of a knowledge source weight , the information level of knowledge source prv information ; The information level prv of the knowledge source information The calculation formula is: in, provider level is the registration level or reputation level index of the knowledge source; provider number is the contribution index of the knowledge source; provider distance is the spatial deviation distance index between the knowledge source and the provided knowledge; The registration level or reputation level index of the knowledge source provider level Use the following calculation formula to get: in, Level(provider) is the registration level or reputation level of the knowledge source; MAX_LEVEL is the maximum registration level or reputation level in all knowledge sources; The higher the registration level or reputation level of the knowledge source, the higher the registration level or reputation level index provider level The bigger; The contribution index of the knowledge source provider number Use the following calculation formula to get: in, number(provider) is the amount of knowledge contributed by the knowledge source; AVERAGE_NUMBER is the average amount of knowledge contributed by all knowledge sources; The more knowledge the knowledge source contributes, the higher the contribution index of the knowledge sourceprovider number The bigger; The spatial deviation distance index between the knowledge source and the provided knowledge provider distance Use the following calculation formula to get: distance(providercity-object) is the Euclidean distance between the knowledge source and the provided knowledge in geographic space; The smaller the Euclidean distance between the knowledge source and the provided knowledge in geographic space, the higher the spatial deviation distance index provider distance The bigger; The hierarchical weight prv of the knowledge source weight Determined by provider tier; Step 1-1-2: Calculate the source reliability Prv of one piece of knowledge among the n pieces of knowledge in the geographic knowledge graph to be measured i ; The source reliability of the piece of knowledge Prv i The calculation formula is: First i =first weight ×first information in, prv weight is the hierarchical weight of the knowledge source; prv information is the information level of the knowledge source; Step 1-1-3: Calculate the source reliability indicator prv , the knowledge source reliability indicator Indicator prv The calculation formula is: in, The geographic knowledge graph to be determined is constructed from n pieces of knowledge; Prv i The reliability of the source of one piece of knowledge.

3. The method for determining the quality of a geographic knowledge graph based on the life cycle theory as claimed in claim 1, characterized in that: In the above steps 1-2, the following steps are specifically included: Step 1-2: Calculate the knowledge source spatiotemporal coverage indicator crg ; The knowledge source spatiotemporal coverage indicator Indicator crg The data stage quality is measured using the spatial and temporal restriction degree of the knowledge source for constructing the geographic knowledge graph to be measured. The higher the spatiotemporal coverage of the knowledge source for constructing the geographic knowledge graph to be measured, the higher the data stage quality; specifically comprising the following steps: Step 1-2-1: Calculate the spatial coverage crg of one piece of knowledge among the n pieces of knowledge constituting the geographic knowledge graph to be measured spatial , time coverage crg temporal ; The spatial coverage crg spatial Use the following calculation formula to get: in, num spatial is the number of spatial knowledge triples in the geographic knowledge graph to be determined; num sp_limited is the number of spatial knowledge triples with spatial area restrictions in the geographic knowledge graph to be determined; The fewer spatial area restrictions there are in the geographic knowledge graph to be measured, the greater the spatial coverage crg spatial The bigger; The time coverage crg temporal Use the following calculation formula to get: in, num temporal is the number of knowledge triples in the geographic knowledge graph to be determined; num tp_limited is the number of spatial knowledge triples with time range restrictions in the geographic knowledge graph to be determined; The fewer time range restrictions there are in the geographic knowledge graph to be measured, the greater the time coverage crg temporal The bigger; Step 1-2-2: Calculate the spatiotemporal coverage Crg of a piece of knowledge source i ; The calculation formula is: Crg i =crg spatial ×crg temporal in, crg spatial is the spatial coverage, which measures the coverage of the knowledge source in space; crg temporal is the temporal coverage, which measures the temporal coverage of the knowledge source; Step 1-2-3: Calculate the knowledge source spatiotemporal coverage indicator crg ; The calculation formula is: in, Cr i It is the spatiotemporal coverage of the knowledge source of one of the n pieces of knowledge constituting the geographic knowledge graph to be measured.

4. The method for determining the quality of a geographic knowledge graph based on the life cycle theory as claimed in claim 1, characterized in that: In the above steps 1-3, The specific steps include: Step 1-3-1, calculate the knowledge medium suitability med of a piece of knowledge in the geographic knowledge graph to be measured suitability , media loss type ; The knowledge medium suitability suitability Use the following calculation formula to get: in, The geographic knowledge graph to be determined is constructed by class knowledge; st i the suitability of the type of medium for one of the categories of knowledge; If such knowledge is transmitted through a suitable medium, then the st i The value is 1.00; If such knowledge is transmitted through an inappropriate medium, the st i The value is 0.50; The media loss type , determined according to the type of knowledge; The knowledge type is structured knowledge, the medium loss is lossless, and the medium lossiness is 1.00; The knowledge type is semi-structured knowledge, the media loss is lossy, and the media lossiness is 0.75; The knowledge type is unstructured knowledge, the media loss is ambiguous, and the media loss is 0.50; Step 1-3-2: Calculate the media reliability Med of a piece of knowledge i ; The formula is Med i =med suitability ×med type in, med suitability is the suitability of the medium of knowledge, which determines whether the knowledge is transmitted through the most appropriate medium; med type The medium lossiness is to determine whether the medium used by the knowledge causes transmission loss to the knowledge; Step 1-3-3: Calculate the knowledge medium reliability indicator med ; The knowledge medium reliability indicator Indicator med The calculation formula is: in, Med i It is the media reliability of one piece of knowledge among the n pieces of knowledge that constitute the geographic knowledge graph to be measured.

5. The method for determining the quality of a geographic knowledge graph based on the life cycle theory as claimed in claim 2, characterized in that: In step 1-1-1, If the provider level is a world or national authoritative data source, the level weight is 1.00; The provider level is a provincial or regional high-confidence data source with a level weight of 0.75; The provider level is a general public data source, and the level weight is 0.50; The provider level is a personal data source, and the level weight is 0.25.

Images