Address Determination Method and System Based on Text Complexity and Node Search
By predicting the final complexity of the address text and determining the initial matching level based on the text complexity algorithm, the problem of insufficient adaptability of address matching strategies in the existing technology is solved, hierarchical and accurate address matching is achieved, and the accuracy and efficiency of address parsing are improved.
Patent Information
- Application Number
- CN202611294517.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-25
- Publication Date
- 2026-11-17
AI Technical Summary
Existing technologies lack dynamic predictive analysis of address text complexity in address resolution, which makes address matching strategies unable to adapt to address texts of different complexities, resulting in insufficient resolution accuracy and efficiency, and easily causing matching deviations.
By obtaining the address text input by the user, the final text complexity is predicted based on the text complexity algorithm, the initial matching level is determined, and node matching search is performed in the address graph node network to achieve hierarchical and accurate address matching.
It improves the accuracy and efficiency of address resolution and reduces the risk of matching errors caused by ambiguous address text or improper hierarchy.
Smart Images

Figure CN123115484A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of data processing technology, and in particular to an address determination method and system based on text complexity and node search. Background Technology
[0002] With the rapid growth of address resolution and geographic information services in logistics, navigation, and smart cities, users and platform operators are increasingly focusing on improving the accuracy and efficiency of address matching through intelligent means. A key technical issue is how to achieve accurate address resolution to reduce the risk of matching errors due to ambiguous address text. Existing technologies typically acquire the address text to be matched from the user input, use fixed-level search or simple keyword matching to perform node matching in the address database, and return results based on unified rules to support address resolution needs. However, existing solutions lack dynamic predictive analysis of address text complexity and a mechanism for intelligently determining the initial matching level based on complexity. This makes it difficult to achieve accurate searches starting from an appropriate level in the address graph node network. Commonly used static or single-level matching strategies cannot adapt to address texts of varying complexity, resulting in insufficient accuracy and efficiency in address resolution. Improper handling of inappropriate levels or fuzzy text can easily lead to matching errors, limiting the overall performance and user experience of the address service system. Therefore, existing technologies have shortcomings that urgently need to be addressed. Summary of the Invention
[0003] The technical problem to be solved by the present invention is to provide an address determination method and system based on text complexity and node search, which can realize hierarchical accurate address matching driven by text complexity, improve the accuracy and efficiency of address parsing, and reduce the risk of matching deviation caused by fuzzy address text or improper hierarchy.
[0004] To address the aforementioned technical problems, the first aspect of this invention discloses an address determination method based on text complexity and node search, the method comprising: Get the address text to be matched from the user input; Based on the text complexity algorithm, predict the final text complexity corresponding to the address text; Based on the final text complexity, determine the initial matching level corresponding to the address text; In a preset address graph node network, starting from the initial matching level, the address text is searched for node matching to obtain the matching address corresponding to the address text.
[0005] As an optional implementation, in the first aspect of the present invention, predicting the final text complexity corresponding to the address text based on the text complexity algorithm includes: The address text and the corresponding user input process data are input into the final text prediction model to obtain multiple output predicted final texts; Based on the text complexity algorithm, the complexity of each predicted final text is determined; Based on the complexity of all the predicted final texts, calculate the complexity of the final text corresponding to the address text.
[0006] As an optional implementation, in a first aspect of the invention, the input process data includes at least one of the following: keyboard operation data, mouse operation data, touch screen operation data, or voice data during the time period in which the user inputs the address text.
[0007] As an optional implementation, in the first aspect of the present invention, determining the complexity of each predicted final text based on the text complexity algorithm includes: Each predicted final text is input into a trained text complexity prediction model to obtain the corresponding complexity; the text complexity prediction model is trained using a training dataset that includes multiple training address texts and corresponding address complexity annotations.
[0008] As an optional implementation, in the first aspect of the invention, calculating the final text complexity corresponding to the address text based on the complexity corresponding to all the predicted final texts includes: Calculate the weighted average of the complexities corresponding to all the predicted final texts to obtain the final text complexity corresponding to the address text; wherein, the calculation weight corresponding to each complexity is the product of a first weight and a second weight; the first weight is inversely proportional to the text length of the corresponding predicted final text; the second weight is directly proportional to the similarity between the corresponding predicted final text and the address text.
[0009] As an optional implementation, in the first aspect of the present invention, determining the initial matching level corresponding to the address text based on the final text complexity includes: Based on the pre-defined correspondence between complexity ranges and levels, multiple candidate matching levels corresponding to the final text complexity are determined. Based on the historical matching data of each candidate matching level, determine the matching degree parameter corresponding to each candidate matching level; The candidate matching level with the highest matching degree parameter is determined as the initial matching level corresponding to the address text.
[0010] As an optional implementation, in the first aspect of the present invention, determining the matching degree parameter corresponding to each candidate matching level based on historical matching data of each candidate matching level includes: For each candidate matching level, obtain multiple historical successfully matched texts corresponding to that candidate matching level; Calculate the average text similarity between each historically successfully matched text and the address text to obtain the matching degree parameter corresponding to that candidate matching level.
[0011] As an optional implementation, in the first aspect of the invention, the node matching search includes the following steps: The address text and the corresponding final text complexity are input into the trained feature extraction network to obtain the corresponding text features; In a preset address graph node network, determine multiple graph nodes corresponding to the initial matching level; Calculate the feature matching degree between the node feature value and the text feature of each graph node; The node address text corresponding to the graph node with the highest feature matching degree is included in the address text set to be sorted; The graph node with the highest feature matching degree is determined, and multiple graph nodes corresponding to the next level of the initial matching level are repeatedly calculated and searched for feature matching degree until the search is completed, resulting in a complete set of address text. Based on the optimal text combination prediction algorithm, the text combination corresponding to the address text set is determined to obtain the matching address corresponding to the address text.
[0012] A second aspect of this invention discloses an address determination system based on text complexity and node search, the system comprising: The acquisition module is used to acquire the address text to be matched input by the user; The prediction module is used to predict the final text complexity corresponding to the address text based on a text complexity algorithm. The determining module is used to determine the initial matching level corresponding to the address text based on the final text complexity; The matching module is used to perform node matching search on the address text starting from the initial matching level in the preset address graph node network to obtain the matching address corresponding to the address text.
[0013] As an optional implementation, in the second aspect of the present invention, the specific method by which the prediction module predicts the final text complexity corresponding to the address text based on a text complexity algorithm includes: The address text and the corresponding user input process data are input into the final text prediction model to obtain multiple output predicted final texts; Based on the text complexity algorithm, the complexity of each predicted final text is determined; Based on the complexity of all the predicted final texts, calculate the complexity of the final text corresponding to the address text.
[0014] As an optional implementation, in a second aspect of the invention, the input process data includes at least one of the following: keyboard operation data, mouse operation data, touch screen operation data, or voice data during the time period in which the user inputs the address text.
[0015] As an optional implementation, in a second aspect of the invention, the prediction module determines the specific method by which it determines the complexity of each predicted final text based on a text complexity algorithm, including: Each predicted final text is input into a trained text complexity prediction model to obtain the corresponding complexity; the text complexity prediction model is trained using a training dataset that includes multiple training address texts and corresponding address complexity annotations.
[0016] As an optional implementation, in a second aspect of the invention, the specific method by which the prediction module calculates the final text complexity corresponding to the address text based on the complexity corresponding to all the predicted final texts includes: Calculate the weighted average of the complexities corresponding to all the predicted final texts to obtain the final text complexity corresponding to the address text; wherein, the calculation weight corresponding to each complexity is the product of a first weight and a second weight; the first weight is inversely proportional to the text length of the corresponding predicted final text; the second weight is directly proportional to the similarity between the corresponding predicted final text and the address text.
[0017] As an optional implementation, in a second aspect of the invention, the determining module determines the specific method for determining the initial matching level corresponding to the address text based on the final text complexity, including: Based on the pre-defined correspondence between complexity ranges and levels, multiple candidate matching levels corresponding to the final text complexity are determined. Based on the historical matching data of each candidate matching level, determine the matching degree parameter corresponding to each candidate matching level; The candidate matching level with the highest matching degree parameter is determined as the initial matching level corresponding to the address text.
[0018] As an optional implementation, in a second aspect of the invention, the determining module determines the matching degree parameter corresponding to each candidate matching level based on historical matching data for each candidate matching level in a specific manner, including: For each candidate matching level, obtain multiple historical successfully matched texts corresponding to that candidate matching level; Calculate the average text similarity between each historically successfully matched text and the address text to obtain the matching degree parameter corresponding to that candidate matching level.
[0019] As an optional implementation, in a second aspect of the invention, the node matching search includes the following steps: The address text and the corresponding final text complexity are input into the trained feature extraction network to obtain the corresponding text features; In a preset address graph node network, determine multiple graph nodes corresponding to the initial matching level; Calculate the feature matching degree between the node feature value and the text feature of each graph node; The node address text corresponding to the graph node with the highest feature matching degree is included in the address text set to be sorted; The graph node with the highest feature matching degree is determined, and multiple graph nodes corresponding to the next level of the initial matching level are repeatedly calculated and searched for feature matching degree until the search is completed, resulting in a complete set of address text. Based on the optimal text combination prediction algorithm, the text combination corresponding to the address text set is determined to obtain the matching address corresponding to the address text.
[0020] A third aspect of this invention discloses another address determination system based on text complexity and node search, the system comprising: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute some or all of the steps in the address determination method based on text complexity and node search disclosed in the first aspect of the present invention.
[0021] The fourth aspect of the present invention discloses a computer storage medium storing computer instructions, which, when invoked, are used to execute some or all of the steps in the address determination method based on text complexity and node search disclosed in the first aspect of the present invention.
[0022] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: This invention obtains the address text to be matched input by the user, predicts the final text complexity based on the text complexity algorithm, determines the initial matching level according to the complexity, and performs node matching search in the address graph node network starting from this level to obtain the matching address. This enables hierarchical and accurate address matching based on text complexity, improves the accuracy and efficiency of address parsing, and reduces the risk of matching deviation caused by ambiguous address text or improper level. Attached Figure Description
[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0024] Figure 1 This is a flowchart illustrating an address determination method based on text complexity and node search disclosed in an embodiment of the present invention.
[0025] Figure 2 This is a schematic diagram of the structure of an address determination system based on text complexity and node search disclosed in an embodiment of the present invention.
[0026] Figure 3 This is a schematic diagram of another address determination system based on text complexity and node search disclosed in an embodiment of the present invention. Detailed Implementation
[0027] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0028] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this invention are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, apparatus, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.
[0029] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.
[0030] This invention discloses an address determination method and system based on text complexity and node search. By acquiring the user-input address text to be matched, the final text complexity is predicted based on a text complexity algorithm. An initial matching level is determined according to the complexity, and node matching search is performed in the address graph node network starting from this level to obtain the matching address. This enables hierarchical and accurate address matching driven by text complexity, improving the accuracy and efficiency of address parsing and reducing the risk of matching deviations caused by ambiguous address text or inappropriate levels. Detailed explanations follow.
[0031] Example 1 Please see Figure 1 , Figure 1 This is a flowchart illustrating an address determination method based on text complexity and node search disclosed in an embodiment of the present invention. Figure 1 The described address determination method based on text complexity and node search can be applied to data processing systems / data processing equipment / data processing servers (including local processing servers or cloud processing servers). Figure 1 As shown, the address determination method based on text complexity and node search can include the following operations: 101. Obtain the address text to be matched input by the user.
[0032] Optionally, the address text to be matched can be the incomplete or missing city / street text entered by the user in the non-standardized input box of the e-commerce order interface, the unstructured logistics address with typos or pinyin abbreviations, or a text phrase containing colloquialisms transcribed by a smart home voice assistant. This invention does not limit the scope of the text.
[0033] 102. Based on the text complexity algorithm, predict the final text complexity corresponding to the address text.
[0034] Optionally, the final text complexity can be a quantitative score scalar representing the severity of missing administrative divisions in the address text, a feature index used to measure the density of noise characters in the text, or a high-dimensional information entropy vector mapped to a nonlinear space. This invention does not limit the specific type of text complexity.
[0035] 103. Determine the initial matching level corresponding to the address text based on the final text complexity.
[0036] Optionally, the initial matching level can be national, provincial, municipal, district / county, township / street, or community / village level; this invention does not impose any limitations.
[0037] 104. In the preset address graph node network, start from the initial matching level to perform node matching search on the address text to obtain the matching address corresponding to the address text.
[0038] Optionally, the address map node network can be a national standard place name address topology map stored in a distributed graph database, a geographic information system (GIS) association network that integrates spatial latitude and longitude grids, or a city multi-level gridded address map constructed based on RDF triples. This invention does not limit the scope of the invention.
[0039] As can be seen, the above-described embodiments of the invention obtain the address text to be matched input by the user, predict the final text complexity based on the text complexity algorithm, determine the initial matching level according to the complexity, and perform node matching search in the address graph node network starting from that level to obtain the matching address. This enables hierarchical and accurate address matching based on text complexity, improves the accuracy and efficiency of address parsing, and reduces the risk of matching deviation caused by ambiguous address text or improper level.
[0040] As an optional embodiment, the step above, predicting the final text complexity corresponding to the address text based on the text complexity algorithm, includes: The address text and the corresponding user input process data are input into the final text prediction model to obtain multiple output predicted final texts; Based on the text complexity algorithm, determine the complexity of each predicted final text; Calculate the complexity of the final text corresponding to the address text based on the complexity of all predicted final texts.
[0041] Optionally, the input process data includes at least one of the following: keyboard operation data, mouse operation data, touch screen operation data, or voice data during the time period when the user is inputting the address text.
[0042] Optionally, the user's input process data may include the time the user holds keys on the soft keyboard, the geometric parameters of the sliding trajectory on the touch screen, the frequency of backspace deletion during the input process, or the background signal-to-noise ratio and acoustic pause characteristics during voice input. This invention does not impose any limitations on these parameters.
[0043] Optionally, the final text prediction model can be a sequence generation large model based on an autoregressive architecture, a streaming text completion network with a bidirectional attention mechanism, or a Markov chain prediction engine combined with a historical association lexicon; this invention does not limit the model.
[0044] Optionally, the predicted final text can be a standardized address string with the province and city prefixes completed, or a standardized geographical entity name after correcting typos; this invention does not impose any limitations on this.
[0045] As can be seen, through the above optional embodiments, multiple predicted final texts are obtained by inputting address text and user input process data into the final text prediction model, and the complexity of all predicted texts is calculated based on the text complexity algorithm to obtain the final text complexity. This achieves accurate complexity assessment based on the fusion of user input behavior and predicted text, improves the dynamism and personalization of complexity prediction, and reduces the risk of complexity deviation caused by relying solely on the current text.
[0046] As an optional embodiment, the step above, determining the complexity of each predicted final text based on a text complexity algorithm, includes: Each predicted final text is input into the trained text complexity prediction model to obtain the corresponding complexity.
[0047] Optionally, the text complexity prediction model is trained using a training dataset that includes multiple training address texts and corresponding address complexity annotations.
[0048] Optionally, the text complexity prediction model can be a fine-tuned bidirectional encoder representation model BERT, a deep long short-term memory network LSTM regressor, or an integrated gradient boosting decision tree XGBoost model; this invention does not limit the model.
[0049] As can be seen, through the above optional embodiments, by obtaining the corresponding complexity from the text complexity prediction model trained on each predicted final text input, the model-based accurate text complexity quantification is achieved, improving the reliability and consistency of complexity calculation, and reducing the risk of inaccurate complexity evaluation due to fixed rules.
[0050] As an optional embodiment, the step above, calculating the final text complexity corresponding to the address text based on the complexity corresponding to all predicted final texts, includes: Calculate the weighted average of the complexities corresponding to all predicted final texts to obtain the final text complexity corresponding to the address text.
[0051] Optionally, the computational weight corresponding to each complexity is the product of the first weight and the second weight.
[0052] Optionally, the first weight is inversely proportional to the text length of the corresponding predicted final text.
[0053] Optionally, the second weight is proportional to the similarity between the corresponding predicted final text and the address text.
[0054] As can be seen, through the above optional embodiments, the final text complexity is obtained by calculating the weighted summation average of all predicted final text complexity, where the weight is the product of a first weight that is inversely proportional to the text length and a second weight that is directly proportional to the text similarity. This achieves a comprehensive evaluation of complexity through multi-factor weighted optimization, improves the balance and adaptability of the final text complexity calculation, and reduces the risk of evaluation bias caused by a single weight.
[0055] As an optional embodiment, the step above, determining the initial matching level corresponding to the address text based on the final text complexity, includes: Based on the pre-defined correspondence between complexity ranges and levels, multiple candidate matching levels corresponding to the final text complexity are determined. Based on the historical matching data of each candidate matching level, determine the matching degree parameter corresponding to each candidate matching level; The candidate matching level with the highest matching degree parameter is determined as the initial matching level corresponding to the address text.
[0056] Optionally, the correspondence between the complexity range and the level can be a threshold hierarchical mapping table pre-fixed in the memory cache, or a complexity ladder boundary dynamically divided by a clustering algorithm. This invention does not limit the scope of the problem.
[0057] Optionally, the candidate matching level can be two or three adjacent administrative division levels that are activated in the matching strategy according to the current complexity score, and this invention does not limit this.
[0058] As can be seen, through the above optional embodiments, multiple candidate matching levels are determined based on a preset complexity range and hierarchical correspondence. The matching degree parameter is calculated based on the historical matching data of each candidate level, and the highest one is selected as the initial matching level. This achieves accurate level selection based on historical matching optimization, improves the accuracy of the address search starting point, and reduces the risk of low subsequent matching efficiency due to improper level selection.
[0059] As an optional embodiment, the step above, determining the matching degree parameter corresponding to each candidate matching level based on historical matching data for each candidate matching level, includes: For each candidate matching level, retrieve multiple historical successfully matched texts corresponding to that candidate matching level; Calculate the average text similarity between each historically successfully matched text and the address text to obtain the matching degree parameter corresponding to that candidate matching level.
[0060] Optionally, the historical successfully matched text can be a standard address sample that has been successfully segmented and delivered at this level within a preset historical period (such as the last 30 days). This invention does not impose any limitations on this.
[0061] Optionally, the text similarity can be the literal overlap calculated using the Levenshtein Distance algorithm, or the bag-of-words intersection-union ratio calculated using the Jaccard similarity coefficient; this invention does not limit the specific text similarity.
[0062] As can be seen, through the above optional embodiments, by obtaining the historical successfully matched text corresponding to each candidate matching level, calculating the average text similarity with the address text to obtain the matching degree parameter, the hierarchical matching quantification based on historical successful cases is realized, the objectivity and reliability of the level selection are improved, and the risk of matching degree deviation caused by the underutilization of historical data is reduced.
[0063] As an optional embodiment, the node matching search in the above steps includes the following steps: The address text and the corresponding final text complexity are input into the trained feature extraction network to obtain the corresponding text features; In the preset address graph node network, determine multiple graph nodes corresponding to the initial matching level; Calculate the feature matching degree between the node feature value and the text feature of each graph node; The node address text corresponding to the graph node with the highest feature matching degree is included in the address text set to be sorted; The graph node with the highest feature matching degree is identified, and multiple graph nodes corresponding to the next level of the initial matching level are identified. The feature matching degree is repeatedly calculated and searched until the search is completed, and a complete set of address text is obtained. Based on the optimal text combination prediction algorithm, the text combination corresponding to the address text set is determined in order to obtain the matching address corresponding to the address text.
[0064] Optionally, the feature extraction network can be a text semantic encoder based on the Transformer architecture, a deep neural network (DNN) feature mapping matrix, or a multimodal joint vectorization layer that incorporates numerical complexity factors; this invention does not limit the type of network.
[0065] Optionally, the graph node can be an entity vertex representing a specific administrative division or physical place name in the graph database. Each vertex is bound to a node feature value that includes the standard full name, alias abbreviation, and spatial latitude and longitude fence. This invention does not impose any limitations on this.
[0066] Optionally, the feature matching degree can be quantified using the inner product score, cosine similarity, or Mahalanobis distance calculated in a low-dimensional dense vector space; this invention does not impose any limitation on this.
[0067] Optionally, the optimal combination prediction algorithm for text can be a conditional random field optimal path search algorithm based on the Viterbi algorithm, a maximum probability address fragment splicing strategy, or a probabilistic graph optimal topology path selection algorithm based on beam search. This invention does not limit the algorithm.
[0068] As can be seen, through the above optional embodiments, text features are obtained by inputting the address text and the final text complexity into the trained feature extraction network. The matching degree between the node feature value and the text feature is calculated in the graph node corresponding to the initial matching level. The matching address is obtained by progressively searching down the level and based on the optimal combination prediction algorithm of the text. This achieves accurate address parsing of hierarchical feature matching and optimal combination, improves the completeness and accuracy of address matching, and reduces the risk of incomplete matching results due to suboptimal search paths.
[0069] Example 2 Please see Figure 2 , Figure 2 This is a schematic diagram of the structure of an address determination system based on text complexity and node search, as disclosed in an embodiment of the present invention. Figure 2 The described address determination system based on text complexity and node search can be applied to data processing systems / data processing equipment / data processing servers (including local processing servers or cloud processing servers). Figure 2 As shown, this address determination system based on text complexity and node search may include: The acquisition module 201 is used to acquire the address text to be matched input by the user.
[0070] The prediction module 202 is used to predict the final text complexity corresponding to the address text based on the text complexity algorithm.
[0071] The determination module 203 is used to determine the initial matching level corresponding to the address text based on the final text complexity.
[0072] The matching module 204 is used to perform node matching search on the address text in the preset address graph node network, starting from the initial matching level, so as to obtain the matching address corresponding to the address text.
[0073] As can be seen, the above-described embodiments of the invention obtain the address text to be matched input by the user, predict the final text complexity based on the text complexity algorithm, determine the initial matching level according to the complexity, and perform node matching search in the address graph node network starting from that level to obtain the matching address. This enables hierarchical and accurate address matching based on text complexity, improves the accuracy and efficiency of address parsing, and reduces the risk of matching deviation caused by ambiguous address text or improper level.
[0074] As an optional embodiment, the prediction module predicts the final text complexity corresponding to the address text using a text complexity algorithm, including the following specific methods: The address text and the corresponding user input process data are input into the final text prediction model to obtain multiple output predicted final texts; Based on the text complexity algorithm, determine the complexity of each predicted final text; Calculate the complexity of the final text corresponding to the address text based on the complexity of all predicted final texts.
[0075] As can be seen, through the above optional embodiments, multiple predicted final texts are obtained by inputting address text and user input process data into the final text prediction model, and the complexity of all predicted texts is calculated based on the text complexity algorithm to obtain the final text complexity. This achieves accurate complexity assessment based on the fusion of user input behavior and predicted text, improves the dynamism and personalization of complexity prediction, and reduces the risk of complexity deviation caused by relying solely on the current text.
[0076] As an optional embodiment, the input process data includes at least one of the following: keyboard operation data, mouse operation data, touch screen operation data, or voice data during the time period in which the user inputs the address text.
[0077] As can be seen, the above optional embodiments limit the data content of the input process data to effectively characterize the features of user input behavior, which facilitates subsequent assistance in achieving accurate prediction of the predicted text and improves the dynamism and personalization of complexity prediction.
[0078] As an optional embodiment, the prediction module determines the specific method for the complexity of each predicted final text based on a text complexity algorithm, including: Each predicted final text is input into the trained text complexity prediction model to obtain the corresponding complexity; the text complexity prediction model is trained using a training dataset that includes multiple training address texts and corresponding address complexity labels.
[0079] As can be seen, through the above optional embodiments, by obtaining the corresponding complexity from the text complexity prediction model trained on each predicted final text input, the model-based accurate text complexity quantification is achieved, improving the reliability and consistency of complexity calculation, and reducing the risk of inaccurate complexity evaluation due to fixed rules.
[0080] As an optional embodiment, the prediction module calculates the complexity of the final text corresponding to the address text based on the complexity of all predicted final texts, including the following specific methods: Calculate the weighted average of the complexities corresponding to all predicted final texts to obtain the final text complexity corresponding to the address text; optionally, the calculation weight corresponding to each complexity is the product of the first weight and the second weight; the first weight is inversely proportional to the text length of the corresponding predicted final text; the second weight is directly proportional to the similarity between the corresponding predicted final text and the address text.
[0081] As can be seen, through the above optional embodiments, the final text complexity is obtained by calculating the weighted summation average of all predicted final text complexity, where the weight is the product of a first weight that is inversely proportional to the text length and a second weight that is directly proportional to the text similarity. This achieves a comprehensive evaluation of complexity through multi-factor weighted optimization, improves the balance and adaptability of the final text complexity calculation, and reduces the risk of evaluation bias caused by a single weight.
[0082] As an optional embodiment, the determining module determines the specific method for determining the initial matching level corresponding to the address text based on the final text complexity, including: Based on the pre-defined correspondence between complexity ranges and levels, multiple candidate matching levels corresponding to the final text complexity are determined. Based on the historical matching data of each candidate matching level, determine the matching degree parameter corresponding to each candidate matching level; The candidate matching level with the highest matching degree parameter is determined as the initial matching level corresponding to the address text.
[0083] As can be seen, through the above optional embodiments, multiple candidate matching levels are determined based on a preset complexity range and hierarchical correspondence. The matching degree parameter is calculated based on the historical matching data of each candidate level, and the highest one is selected as the initial matching level. This achieves accurate level selection based on historical matching optimization, improves the accuracy of the address search starting point, and reduces the risk of low subsequent matching efficiency due to improper level selection.
[0084] As an optional embodiment, the determining module determines the specific method for the matching degree parameter corresponding to each candidate matching level based on the historical matching data of each candidate matching level, including: For each candidate matching level, retrieve multiple historical successfully matched texts corresponding to that candidate matching level; Calculate the average text similarity between each historically successfully matched text and the address text to obtain the matching degree parameter corresponding to that candidate matching level.
[0085] As can be seen, through the above optional embodiments, by obtaining the historical successfully matched text corresponding to each candidate matching level, calculating the average text similarity with the address text to obtain the matching degree parameter, the hierarchical matching quantification based on historical successful cases is realized, the objectivity and reliability of the level selection are improved, and the risk of matching degree deviation caused by the underutilization of historical data is reduced.
[0086] As an optional embodiment, node matching search includes the following steps: The address text and the corresponding final text complexity are input into the trained feature extraction network to obtain the corresponding text features; In the preset address graph node network, determine multiple graph nodes corresponding to the initial matching level; Calculate the feature matching degree between the node feature value and the text feature of each graph node; The node address text corresponding to the graph node with the highest feature matching degree is included in the address text set to be sorted; The graph node with the highest feature matching degree is identified, and multiple graph nodes corresponding to the next level of the initial matching level are identified. The feature matching degree is repeatedly calculated and searched until the search is completed, and a complete set of address text is obtained. Based on the optimal text combination prediction algorithm, the text combination corresponding to the address text set is determined in order to obtain the matching address corresponding to the address text.
[0087] As can be seen, through the above optional embodiments, text features are obtained by inputting the address text and the final text complexity into the trained feature extraction network. The matching degree between the node feature value and the text feature is calculated in the graph node corresponding to the initial matching level. The matching address is obtained by progressively searching down the level and based on the optimal combination prediction algorithm of the text. This achieves accurate address parsing of hierarchical feature matching and optimal combination, improves the completeness and accuracy of address matching, and reduces the risk of incomplete matching results due to suboptimal search paths.
[0088] Example 3 Please see Figure 3 , Figure 3 This is another address determination system based on text complexity and node search disclosed in the embodiments of the present invention. Figure 3 The described address determination system based on text complexity and node search is applied in data processing systems / data processing equipment / data processing servers (wherein, the server includes local processing servers or cloud processing servers). Figure 3 As shown, this address determination system based on text complexity and node search may include: Memory 301 storing executable program code; Processor 302 coupled to memory 301; The processor 302 calls the executable program code stored in the memory 301 to execute the steps of the address determination method based on text complexity and node search described in Embodiment 1.
[0089] Example 4 This invention discloses a computer read storage medium that stores a computer program for electronic data interchange, wherein the computer program causes a computer to execute the steps of the address determination method based on text complexity and node search described in Embodiment 1.
[0090] Example 5 This invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to perform the steps of the address determination method based on text complexity and node search described in Embodiment 1.
[0091] The foregoing has described specific embodiments of this specification; other embodiments are within the scope of the appended claims. In some cases, the actions or steps described in the claims may be performed in a different order than those shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily have to follow the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0092] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.
[0093] For ease of description, the above devices are described in terms of function, divided into various units. Of course, in implementing this specification, the functions of each unit can be implemented in one or more software and / or hardware.
[0094] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, the embodiments of this specification can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the embodiments of this specification can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0095] This specification is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0096] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0097] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0098] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0099] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.
[0100] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.
[0101] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
[0102] This specification can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This specification can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.
[0103] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.
[0104] Finally, it should be noted that the address determination method and system based on text complexity and node search disclosed in the embodiments of the present invention are merely preferred embodiments of the present invention and are only used to illustrate the technical solutions of the present invention, not to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. An address determination method based on text complexity and node search, characterized in that, The method includes: Get the address text to be matched from the user input; Based on the text complexity algorithm, predict the final text complexity corresponding to the address text; Based on the final text complexity, determine the initial matching level corresponding to the address text; In a preset address graph node network, starting from the initial matching level, the address text is searched for node matching to obtain the matching address corresponding to the address text.
2. The address determination method based on text complexity and node search according to claim 1, characterized in that, The text complexity-based algorithm for predicting the final text complexity corresponding to the address text includes: The address text and the corresponding user input process data are input into the final text prediction model to obtain multiple output predicted final texts; Based on the text complexity algorithm, the complexity of each predicted final text is determined; Based on the complexity of all the predicted final texts, calculate the complexity of the final text corresponding to the address text.
3. The address determination method based on text complexity and node search according to claim 2, characterized in that, The input process data includes at least one of the following: keyboard operation data, mouse operation data, touch screen operation data, or voice data during the time period when the user inputs the address text.
4. The address determination method based on text complexity and node search according to claim 2, characterized in that, The text complexity algorithm determines the complexity of each predicted final text, including: Each predicted final text is input into a trained text complexity prediction model to obtain the corresponding complexity; the text complexity prediction model is trained using a training dataset that includes multiple training address texts and corresponding address complexity annotations.
5. The address determination method based on text complexity and node search according to claim 2, characterized in that, The step of calculating the final text complexity corresponding to the address text based on the complexity of all predicted final texts includes: Calculate the weighted average of the complexities corresponding to all the predicted final texts to obtain the final text complexity corresponding to the address text; wherein, the calculation weight corresponding to each complexity is the product of a first weight and a second weight; the first weight is inversely proportional to the text length of the corresponding predicted final text; the second weight is directly proportional to the similarity between the corresponding predicted final text and the address text.
6. The address determination method based on text complexity and node search according to claim 1, characterized in that, The step of determining the initial matching level corresponding to the address text based on the final text complexity includes: Based on the pre-defined correspondence between complexity ranges and levels, multiple candidate matching levels corresponding to the final text complexity are determined. Based on the historical matching data of each candidate matching level, determine the matching degree parameter corresponding to each candidate matching level; The candidate matching level with the highest matching degree parameter is determined as the initial matching level corresponding to the address text.
7. The address determination method based on text complexity and node search according to claim 6, characterized in that, The step of determining the matching degree parameter corresponding to each candidate matching level based on historical matching data of each candidate matching level includes: For each candidate matching level, obtain multiple historical successfully matched texts corresponding to that candidate matching level; Calculate the average text similarity between each historically successfully matched text and the address text to obtain the matching degree parameter corresponding to that candidate matching level.
8. The address determination method based on text complexity and node search according to claim 1, characterized in that, The node matching search includes the following steps: The address text and the corresponding final text complexity are input into the trained feature extraction network to obtain the corresponding text features; In a preset address graph node network, determine multiple graph nodes corresponding to the initial matching level; Calculate the feature matching degree between the node feature value and the text feature of each graph node; The node address text corresponding to the graph node with the highest feature matching degree is included in the address text set to be sorted; The graph node with the highest feature matching degree is determined, and multiple graph nodes corresponding to the next level of the initial matching level are repeatedly calculated and searched for feature matching degree until the search is completed, resulting in a complete set of address text. Based on the optimal text combination prediction algorithm, the text combination corresponding to the address text set is determined to obtain the matching address corresponding to the address text.
9. An address determination system based on text complexity and node search, characterized in that, The system includes: The acquisition module is used to acquire the address text to be matched input by the user; The prediction module is used to predict the final text complexity corresponding to the address text based on a text complexity algorithm. The determining module is used to determine the initial matching level corresponding to the address text based on the final text complexity; The matching module is used to perform node matching search on the address text starting from the initial matching level in the preset address graph node network to obtain the matching address corresponding to the address text.
10. An address determination system based on text complexity and node search, characterized in that, The system includes: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the address determination method based on text complexity and node search as described in any one of claims 1-8.
