Information processing method, device and equipment and computer readable storage medium

Abstract

The embodiment of the invention discloses an information processing method and device, equipment and a computer readable storage medium, and the method comprises the steps: determining a first similarity between to-be-analyzed texts based on the text features of the to-be-analyzed texts, and dividing the to-be-analyzed texts into a plurality of text sets based on the first similarity; a plurality of semantic similarity models are adopted to determine the second similarity between the first text and each second text in each second-class text set in the text set, and each second-class text set comprises the first text and a plurality of second texts; processing a plurality of second texts in each second type of text set based on the second similarity to obtain an optimized text set; and on the basis of the first type of text set in the text set, the optimized text set and the target hierarchical information, generating the target text set with the hierarchical relationship for the to-be-analyzed text, so that not only is the accuracy of determining the similarity between the texts improved, but also the efficiency of retrieving the problem input by the user in the target text set is improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to an information processing method, device, equipment and computer-readable storage medium. Background Art

[0002] Typically, the knowledge base of a question-answering system consists of multiple documents, and a question and answer (Q&A) pair is considered as one document; text matching technology can be used to recommend similar questions and answers to similar questions to the question entered by the user from the knowledge base.

[0003] When organizing a knowledge base, a single semantic similarity clustering method is usually used to identify similar questions in the knowledge base as a problem cluster, and then the constructed multiple problem clusters are stored in the knowledge base in a parallel storage structure; however, in the process of determining problem clusters in related technologies, the single semantic similarity clustering method usually obtains semantic deviations of the questions, thereby identifying dissimilar questions as a problem cluster, resulting in low accuracy in determining problem clusters, and multiple problem clusters are stored in a parallel structure, resulting in low efficiency in retrieving user-entered questions in the knowledge base. Summary of the invention

[0004] In order to solve the above-mentioned technical problems, the embodiments of the present application hope to provide an information processing method, apparatus, device and computer-readable storage medium, which can solve the problems of low accuracy in determining question clusters and low efficiency in retrieving user-input questions in the knowledge base in the related technology, and not only improve the accuracy of determining the similarity between texts, but also improve the efficiency of retrieving user-input questions in the target text set.

[0005] The technical solution of this application is implemented as follows:

[0006] An information processing method, the method comprising:

[0007] Determine a first similarity between the texts to be analyzed based on the text features of the texts to be analyzed, and divide the plurality of texts to be analyzed into a plurality of text sets based on the first similarity; wherein the plurality of text sets include a first type of text set having one text to be analyzed and a second type of text set having a plurality of texts to be analyzed;

[0008] Using multiple semantic similarity models, determine the second similarity between the first text and each second text in each second-category text set; wherein each second-category text set includes the first text and multiple second texts;

[0009] Processing the plurality of second texts in each of the second-category text sets based on the second similarity to obtain an optimized text set;

[0010] Based on the first type of text set, the optimized text set and the target hierarchical information, a target text set with a hierarchical relationship for the text to be analyzed is generated.

[0011] In the above scheme, determining the first similarity between the texts to be analyzed based on the text features of the texts to be analyzed, and dividing the plurality of texts to be analyzed into a plurality of text sets based on the first similarity, comprises:

[0012] Determine the number of identical characters between the texts to be analyzed;

[0013] Determine the difference between the total number of characters in the text to be analyzed and the number of the same characters;

[0014] Calculating the difference and the number of identical characters to obtain the first similarity;

[0015] Based on the first similarity and the similarity threshold, the multiple texts to be analyzed are divided into the multiple text sets.

[0016] In the above solution, the dividing the plurality of texts to be analyzed into the plurality of text sets based on the first similarity and the similarity threshold comprises:

[0017] When the maximum similarity among the first similarities meets a first threshold, dividing the plurality of texts to be analyzed into the first category of text sets based on a first keyword;

[0018] In the case where the maximum similarity satisfies the second threshold, determining the text to be analyzed that satisfies the third threshold as the first text set;

[0019] In the case where the maximum similarity satisfies a fourth threshold, based on the number of texts to be analyzed that satisfy a fifth threshold, dividing the texts to be analyzed into a second text set;

[0020] In a case where the maximum similarity satisfies a sixth threshold, sorting the text to be analyzed based on the first similarity to obtain a first sorted text;

[0021] The M texts to be analyzed in the first sorted texts are determined as a third text set; wherein the second type of text set includes the first text set, the second text set and the third text set.

[0022] In the above solution, the dividing the plurality of texts to be analyzed into the first type of text set based on the first keyword includes:

[0023] In the case that the first keyword does not exist in the text to be analyzed, dividing the text to be analyzed into a first type of text set;

[0024] In the case where the first keyword exists in the text to be analyzed, deleting the first keyword in the text to be analyzed, and determining a third similarity between the texts after the keyword is deleted;

[0025] When the maximum similarity among the third similarities meets the first threshold, the text to be analyzed corresponding to the text after the keyword is deleted is divided into a first type of text set.

[0026] In the above solution, when the maximum similarity satisfies the fourth threshold, the text to be analyzed is divided into a second text set based on the number of texts to be analyzed that meet the fifth threshold, including:

[0027] When the number of the to-be-analyzed texts satisfying the fifth threshold is greater than or equal to the first number, dividing the to-be-analyzed texts satisfying the fifth threshold into the second text set;

[0028] In a case where the number of the to-be-analyzed texts satisfying the fifth threshold is less than the first number, sorting the to-be-analyzed texts based on the first similarity to obtain second sorted texts;

[0029] The first number of to-be-analyzed texts in the second sorted texts are divided into the second text set.

[0030] In the above solution, the step of processing the plurality of second texts in each second type text set based on the second similarity to obtain an optimized text set includes:

[0031] For a second type of text set whose target similarity in the second similarity satisfies a seventh threshold, determining that the first text and each of the second texts in the second type of text set are each a first optimized text set;

[0032] For a second type of text set whose target similarity in the second similarity meets an eighth threshold, a second text in the second type of text set that meets the eighth threshold and the first text are determined as a second optimized text set; wherein the optimized text set includes the first optimized text set and the second optimized text set.

[0033] In the above solution, the information processing method further includes:

[0034] For the second type of text set whose target similarity in the second similarity does not satisfy the seventh threshold and the eighth threshold, determining a fourth similarity between each to-be-processed text in the second type of text set that does not satisfy the seventh threshold and the eighth threshold and a text in the second optimized text set;

[0035] Determine a first target text in each of the second optimized text sets;

[0036] Based on the fourth similarity and the first target text, the to-be-processed text is divided into the first optimized text set and the second optimized text set.

[0037] In the above solution, the step of determining the first target text in each of the second optimized text sets includes:

[0038] Determining a density parameter and a distance parameter of the first text and each of the second texts in the second optimized text set;

[0039] The first target text is determined based on the density parameter and the distance parameter.

[0040] In the above solution, the step of dividing the to-be-processed text into the first optimized text set and the second optimized text set based on the fourth similarity and the first target text includes:

[0041] In the case where there are P fourth similarities among the fourth similarities that are greater than a ninth threshold, dividing the to-be-processed text into the second optimized text set;

[0042] In the case that there are not P fourth similarities among the fourth similarities that are greater than the ninth threshold, the to-be-processed text is divided into the first optimized text set and the second optimized text set based on the first target text.

[0043] In the above solution, the step of dividing the to-be-processed text into the first optimized text set and the second optimized text set based on the first target text includes:

[0044] Determining a fifth similarity between each of the to-be-processed texts and each of the first target texts;

[0045] When the maximum similarity among the fifth similarities is less than or equal to the second similarity corresponding to the first target text, dividing the to-be-processed text into the second optimized text set;

[0046] When the maximum similarity among the fifth similarities is greater than the second similarity corresponding to the first target text, the text to be processed is divided into the first optimized text set.

[0047] In the above solution, the information processing method further includes:

[0048] Determine a total number of first occurrences of text in the first type of text set and a total number of second occurrences of text in the optimized text set;

[0049] The first total number of occurrences is marked in the first type of text set, and the second total number of occurrences is marked in the optimized text set to obtain an updated text set.

[0050] In the above solution, the information processing method further includes:

[0051] For a third text in the updated text set whose total number of occurrences is less than the first number threshold, determine a fourth text at the same level as the third text whose total number of occurrences meets the second number threshold; wherein the updated text set includes the third text and the fourth text;

[0052] A fifth text is obtained based on the second keyword in the third text and the characters other than the third keyword in the fourth text.

[0053] In the above solution, the information processing method further includes:

[0054] Receive query text input by the user;

[0055] Acquire a second target text from the target text set, the matching degree of which satisfies a first matching degree threshold with the query text, and display response information of the second target text;

[0056] In the case that the second target text does not exist in the target text set, a third target text whose matching degree with the query text satisfies a second matching degree threshold is determined from the target text set, and response information of the third target text is displayed.

[0057] An information processing device, comprising:

[0058] A first processing unit is used to determine a first similarity between the texts to be analyzed based on text features of the texts to be analyzed, and divide the plurality of texts to be analyzed into a plurality of text sets based on the first similarity; wherein the plurality of text sets include a first type of text set having one text to be analyzed and a second type of text set having a plurality of texts to be analyzed;

[0059] A determination unit, configured to determine a second similarity between the first text and each second text in each of the second-category text sets by using a plurality of semantic similarity models; wherein each of the second-category text sets includes the first text and a plurality of the second texts;

[0060] A second processing unit, configured to process the plurality of second texts in each of the second-category text sets based on the second similarity to obtain an optimized text set;

[0061] A generating unit is used to generate a target text set with a hierarchical relationship for the text to be analyzed based on the first type of text set, the optimized text set and the target hierarchical information.

[0062] An information processing device, the device comprising: a processor, a memory and a communication bus;

[0063] The communication bus is used to realize the communication connection between the processor and the memory;

[0064] The processor is used to execute the information processing program stored in the memory to implement the steps of the above-mentioned information processing method.

[0065] A computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of the above-mentioned information processing method.

[0066] The information processing method, apparatus, device and computer-readable storage medium provided in the embodiments of the present application first determine the first similarity between the texts to be analyzed based on the text features of the texts to be analyzed, and divide the multiple texts to be analyzed into multiple text sets based on the first similarity, and the multiple text sets include a first-category text set with one text to be analyzed and a second-category text set with multiple texts to be analyzed, and then use multiple semantic similarity models to determine the second similarity between the first text and each second text in each second-category text set, and each second-category text set includes the first text and multiple second texts, and then process the multiple second texts in each second-category text set based on the second similarity to obtain an optimized text set. This set, then based on the first type of text set, the optimized text set and the target hierarchical information, generates a target text set with a hierarchical relationship for the text to be analyzed. In this way, firstly, based on the text features, the first type of text set with only one text to be analyzed and the second type of text set with multiple texts to be analyzed are determined, and then only the second type of text set with multiple texts to be analyzed is processed again, which reduces the amount of processed data, and when the second type of text set is processed again, multiple semantic similarity models are used to determine the similarity between texts, thereby improving the accuracy of determining the similarity between texts, and then a target text set with a hierarchical relationship is constructed, which improves the efficiency of retrieving user input questions in the target text set. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 A flowchart of an information processing method provided in an embodiment of the present application;

[0068] Figure 2 A flowchart of another information processing method provided in an embodiment of the present application;

[0069] Figure 3 A schematic diagram of a target text set with a hierarchical relationship in an information processing method provided in an embodiment of the present application;

[0070] Figure 4 A schematic diagram of an intelligent question-answering system in an information processing method provided in an embodiment of the present application;

[0071] Figure 5 A flowchart of another information processing method provided in an embodiment of the present application;

[0072] Figure 6 A schematic diagram of the structure of an information processing device provided in an embodiment of the present application;

[0073] Figure 7 A schematic diagram of the structure of an information processing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0074] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0075] It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0076] The present application provides an information processing method, which can be applied to an information processing device. Figure 1 As shown, the method comprises the following steps:

[0077] Step 101: determine a first similarity between texts to be analyzed based on text features of the texts to be analyzed, and divide a plurality of texts to be analyzed into a plurality of text sets based on the first similarity.

[0078] The multiple text sets include a first type of text set having one text to be analyzed and a second type of text set having multiple texts to be analyzed.

[0079] In an embodiment of the present application, text features may refer to features at the character level; the first similarity may refer to the similarity between any two texts to be analyzed; the first category text set may refer to a text set with only one text to be analyzed, the second category text set may refer to a text set with multiple texts to be analyzed, and the second category text set may include multiple text sets, namely a first text set, a second text set and a third text set; the text features of the texts to be analyzed may be determined first, and then the first similarity between the texts to be analyzed may be determined based on the text features, and after determining the first similarity between the texts to be analyzed, the multiple texts to be analyzed may be divided according to the first similarity to obtain multiple first category text sets and multiple second category text sets; in a feasible implementation method, the text to be analyzed may refer to a question in a knowledge base.

[0080] Step 102: Use multiple semantic similarity models to determine the second similarity between the first text and each second text in each second category text set.

[0081] Each second-category text set includes a first text and a plurality of second texts.

[0082] In an embodiment of the present application, a semantic similarity model may refer to a model that can extract semantic features and determine the similarity of texts based on the semantic features; the second similarity may refer to the similarity between the first text and each second text in each second-category text set, and each second similarity is composed of multiple similarities determined by multiple semantic similarity models, and there are multiple semantic similarity models, and there are correspondingly many similarities in the second similarity, that is, each semantic similarity model can determine one similarity in the second similarity; for each second-category text set, multiple semantic similarity models can be used to process the first text and each second text in the second-category text set to obtain multiple second similarities between the first text and each second text; it should be noted that the use of multiple semantic similarity models can comprehensively consider the results of multiple models, thereby improving the accuracy of text segmentation.

[0083] In a feasible implementation, the first text may refer to the main question, and the second text may refer to a question other than the main question in the text set; the multiple semantic similarity models may be two semantic similarity models, may be three semantic similarity models, and of course, may be other numbers of semantic similarity models, and when the multiple semantic similarity models may be two semantic similarity models, the two semantic similarity models are an enhanced sequential inference model (ESIM) and a link (Linkage), and the first text is "How to satisfy the desire of diabetics to drink porridge and eat steamed buns without soaring blood sugar?", and the second text is "Can't diabetics drink porridge?", the second similarity between the first text and the second text determined by ESIM and Linkage is (0.52, 0.68); when the first text is "At what age is diabetes most likely to occur?" and the second text is "At what age is diabetes most likely to occur?", the second similarity between the first text and the second text determined by ESIM and Linkage is (0.87, 0.61).

[0084] Step 103: Process the multiple second texts in each second type text set based on the second similarity to obtain an optimized text set.

[0085] In an embodiment of the present application, the optimized text set may refer to the one obtained after optimizing the second text in the second category text set; for each second category text set, the multiple second texts in the second category text set may be optimized based on the second similarity between the first text in the second category text set and each second text and multiple similarity thresholds to obtain the optimized text set; it should be noted that the multiple similarity thresholds may be determined based on historical experimental data.

[0086] Step 104: Generate a target text set with a hierarchical relationship for the text to be analyzed based on the first type of text set, the optimized text set and the target hierarchical information.

[0087] In an embodiment of the present application, the texts to be analyzed in the target text set are stored according to a hierarchical relationship; the first category text set and the optimized text set can be respectively mapped in the target hierarchical information based on the target hierarchical information, thereby generating a target text set with a hierarchical relationship; it should be noted that the target hierarchical information can be predetermined based on historical experience; in a feasible implementation method, the target hierarchical information can refer to a preset hierarchical template.

[0088] The information processing method provided in the embodiment of the present application first determines a first type of text set having only one text to be analyzed and a second type of text set having multiple texts to be analyzed based on text features, and then only processes the second type of text set having multiple texts to be analyzed, thereby reducing the amount of data to be processed. When the second type of text set is processed again, multiple semantic similarity models are used to determine the similarity between texts, thereby improving the accuracy of determining the similarity between texts. Subsequently, a target text set with a hierarchical relationship is constructed, thereby improving the efficiency of retrieving user input questions in the target text set.

[0089] Based on the above embodiments, the embodiments of the present application provide an information processing method, referring to Figure 2 As shown, the method comprises the following steps:

[0090] Step 201: The information processing device determines the number of identical characters between the texts to be analyzed.

[0091] In the embodiment of the present application, for any text to be analyzed S a and the text to be analyzed S b , when determining S a and S b When the number of characters between a and S b Analyze the characters in S a and S b The same characters in S are counted and the number of the same characters is obtained.a and S b The number of identical characters between them.

[0092] In a feasible implementation, when the text S to be analyzed a is "Will nervousness affect blood sugar?", the text S to be analyzed b is "Will emotional stress cause blood sugar to rise?", then the identical characters in S a and S b are {emotion, tension, will, blood, sugar,?}, and the number of identical characters is 7.

[0093] Step 202: The information processing device determines the difference between the total number of characters and the number of identical characters between the texts to be analyzed.

[0094] In the embodiments of the present application, for any text S to be analyzed a and the text S to be analyzed b , when it is necessary to determine the total number of characters between S a and S b , all characters in S a and S b can be counted to obtain the total number of characters in S a and S b . Then, the subtraction operation can be performed on the total number of characters in S a and S b and the number of identical characters between S a and S b to obtain the operation result (i.e., the difference).

[0095] In a feasible implementation, when the text S to be analyzed a is "Will nervousness affect blood sugar?", the text S to be analyzed b is "Will emotional stress cause blood sugar to rise?", then the total characters between S a and S b are {heart, emotion, tension, will, affect, blood, sugar,?, emotion, mood, tension, will, cause, blood, sugar, rise,?}, count the characters in S a and S b to obtain the total number of characters between S a and S b is 23, and the number of identical characters between S a and S b is 7, then the difference is obtained as 16.

[0096] Step 203: The information processing device performs an operation on the difference and the number of identical characters to obtain the first similarity.

[0097] In the embodiment of the present application, the number of identical characters and the difference can be divided to obtain the first similarity; it should be noted that the value range of the first similarity is [0, 1]. a and the text to be analyzed S b When they are completely consistent, the first similarity is 1; when the text to be analyzed S a and the text to be analyzed S b When there is no consistency, the first similarity is 0.

[0098] In a feasible implementation, the text to be analyzed S a For "Will stress affect blood sugar?", the text to be analyzed is S b is "Will emotional tension cause blood sugar to rise?", the difference is 16, and the number of identical characters is 7. Then the text to be analyzed S can be calculated by division operation. a and the text to be analyzed S b The first similarity between them is 0.4375.

[0099] Step 204: The information processing device divides the multiple texts to be analyzed into multiple text sets based on the first similarity and the similarity threshold.

[0100] In an embodiment of the present application, multiple text sets may include a first category text set, a first text set, a second text set, and a third text set; a maximum similarity may be first determined from multiple first similarities, and then the maximum similarity may be compared with a similarity threshold, and based on the comparison result, multiple texts to be analyzed may be divided into the first category text set, the first text set in the second category text set, the second text set in the second category text set, and the third text set in the second category text set; it should be noted that the similarity threshold may be determined based on historical experimental data.

[0101] It should be noted that step 204 can be implemented in the following ways:

[0102] Step 204a1: When the maximum similarity among the first similarities satisfies a first threshold, the information processing device divides the plurality of texts to be analyzed into a first category of text sets based on the first keyword.

[0103] In an embodiment of the present application, the maximum similarity among the first similarities satisfying the first threshold may mean that the maximum similarity is less than the first threshold; when the maximum similarity among the first similarities is less than the first threshold, it can be determined whether the first keyword exists in the text to be analyzed, and then the multiple texts to be analyzed are divided into a first category of text sets based on the judgment result.

[0104] In a feasible implementation, the first threshold can be set to 0.25, and the first keyword can refer to the name of the disease; when calculating the first similarity between the text to be analyzed "How long does it take for milky triglycerides to recover" and the text to be analyzed "Can you eat corn and sweet potatoes if you have high triglycerides", "Do you need to take medicine if your triglycerides are 4.56 and your blood sugar is 6.25", ... and "Can you take Centrum tablets if you have diabetes? Can you take them if you smoke?", and "How long does it take for milky triglycerides to recover" and "Can you eat corn and sweet potatoes if you have high triglycerides" are similar to " The first similarity between "how long does it take for triglyceride to recover from milky eating" and "do I need to take medicine if my triglyceride level is 4.56 and my blood sugar level is 6.25" is 0.172, and the first similarity between "how long does it take for triglyceride to recover from milky eating" and "can I take Centrum tablets if I have diabetes and I smoke?" is 0.133. If the first similarity of the texts to be analyzed meets the first threshold of 0.25, the multiple texts to be analyzed are divided into the first category of text sets based on the disease names.

[0105] Step 204a2: When the maximum similarity satisfies the second threshold, the information processing device determines that the text to be analyzed that satisfies the third threshold is the first text set.

[0106] In the embodiment of the present application, the maximum similarity in the first similarities satisfying the second threshold may mean that the maximum similarity is within the second threshold, and satisfying the third threshold may mean that the first similarity is greater than the third threshold.

[0107] In a feasible implementation, the second threshold may refer to [0.25, 0.5], and the third threshold may refer to 0.1. Then, when the maximum similarity is in the interval [0.25, 0.5], the text to be analyzed with a first similarity greater than 0.1 is determined to be the first text set; when calculating the first similarity between the text to be analyzed "Will mood tension affect blood sugar" and the text to be analyzed "Will anemia affect high blood sugar", "Will mood tension cause high blood sugar?" and "Will hypoglycemia cause diabetes", and the first similarity between "Will mood tension affect blood sugar" and "Will anemia affect high blood sugar" The similarity is 0.5, the first similarity between "Will stress affect blood sugar" and "Will stress cause high blood sugar?" is 0.437, the first similarity between "Will stress affect blood sugar" and "Will hypoglycemia cause diabetes?" is 0.266, and the maximum similarity 0.5 among the first similarities of the texts to be analyzed is within the second threshold [0.25, 0.5], then the texts to be analyzed "Will stress affect blood sugar", "Will anemia affect high blood sugar", "Will stress cause high blood sugar?" and "Will hypoglycemia cause diabetes" are determined as the first text set.

[0108] Step 204a3: When the maximum similarity satisfies the fourth threshold, the information processing device divides the texts to be analyzed into a second text set based on the number of texts to be analyzed that meet the fifth threshold.

[0109] In an embodiment of the present application, the maximum similarity in the first similarity satisfying the fourth threshold may mean that the maximum similarity is within the fourth threshold, and satisfying the fifth threshold may mean that the first similarity is greater than the fifth threshold; when the maximum similarity is within the fourth threshold, the number of texts to be analyzed whose first similarity is greater than the fifth threshold can be counted, and the texts to be analyzed can be divided into a second text set based on the statistical results.

[0110] In a feasible implementation, the fourth threshold may refer to (0.5, 0.8], and the fifth threshold may refer to 0.5. Then, when the maximum similarity is within the interval of (0.5, 0.8], the text to be analyzed is determined as the second text set based on the number of texts to be analyzed whose first similarity is greater than 0.5; when the first similarity between the text to be analyzed "At what age is diabetes most likely to occur?" and the text to be analyzed "What group of people is most likely to get diabetes?", "At what age does diabetes occur?", ..., "What kind of sugar should diabetes eat", and "What kind of sugar should diabetes eat?" is calculated, the text to be analyzed is determined as the second text set based on the number of texts to be analyzed whose first similarity is greater than 0.5. The first similarity between "at what age is diabetes most likely to occur" and "which group of people is most likely to get diabetes" is 0.75, the first similarity between "at what age is diabetes most likely to occur" and "at what age does diabetes occur?" is 0.643, the first similarity between "at what age is diabetes most likely to occur" and "what kind of sugar should diabetes patients eat" is 0.417, and the maximum similarity of 0.75 among the first similarities of the texts to be analyzed meets the fourth threshold (0.5, 0.8], then the texts to be analyzed can be determined as the second text set according to the number of texts to be analyzed whose first similarities are greater than 0.5.

[0111] Step 204a4: When the maximum similarity satisfies the sixth threshold, the information processing device sorts the texts to be analyzed based on the first similarity to obtain first sorted texts.

[0112] In an embodiment of the present application, the maximum similarity in the first similarity satisfies the sixth threshold value, which may refer to the maximum similarity being greater than the sixth threshold value, and the first sorted text may refer to the text obtained after sorting the text to be analyzed in descending order according to the first similarity; when the maximum similarity is greater than the sixth threshold value, all the texts to be analyzed can be sorted in descending order according to the first similarity to obtain the sorted text to be analyzed (i.e., the first sorted text).

[0113] In a feasible implementation, the sixth threshold value may refer to 0.8; when calculating the first similarity between the text to be analyzed "What will happen if insulin is too high" and the text to be analyzed "What will cause insulin to be too high", "What happens if insulin is too high", ..., "What is the problem with insulin overflow?", and the first similarity between "What will happen if insulin is too high" and "What will cause insulin to be too high" is 1.0, the first similarity between "What will happen if insulin is too high" and "What will happen if insulin is too high" is 0.461, the first similarity between "What will happen if insulin is too high" and "What is the problem with insulin overflow?" is 0.3, and the maximum similarity 1.0 among the first similarities of the text to be analyzed satisfies the sixth threshold value 0.8, then the text to be analyzed can be sorted according to the first similarity to obtain the first sorted text, namely, what will happen if insulin is too high, what will cause insulin to be too high, what will happen if insulin is too high, ... and what is the problem with insulin overflow? .

[0114] Step 204a5: The information processing device determines the M texts to be analyzed in the first sorted texts as a third text set.

[0115] In an embodiment of the present application, the first M texts to be analyzed can be obtained from the first sorted texts in descending order of the first similarity, and the M texts to be analyzed can be determined as the third text set; in a feasible implementation method, M can be set to 2000.

[0116] Step 205: The information processing device uses multiple semantic similarity models to determine the second similarity between the first text and each second text in each second type of text set.

[0117] Each second-category text set includes a first text and a plurality of second texts.

[0118] Step 206 : For each second-category text set, the information processing device determines that the first text and each second text in the second-category text set are each a first optimized text set when the target similarity in the second similarity satisfies a seventh threshold.

[0119] Among them, the target similarity in the second similarity may refer to the similarity corresponding to the second text with the greatest similarity to the first text determined by using multiple semantic similarity models, and the number of similarities in the target similarity is the same as the number of semantic similarity models adopted, that is, each semantic similarity model can determine one similarity in the target similarity; the seventh threshold may be determined based on historical experimental data; the target similarity in the second similarity satisfying the seventh threshold may mean that each similarity in the target similarity is less than or equal to the seventh threshold.

[0120] In an embodiment of the present application, for each second-category text set, when the target similarity in the second similarity is less than or equal to the seventh threshold, it is directly determined that neither the first text nor the second text in the second-category text set has a clusterable text, and the second text and the second text in the second-category text set are separately grouped into a set (i.e., a cluster) and marked as clustered and do not participate in subsequent processing; it should be noted that each first optimized text set has one text; in a feasible implementation method, when the second texts in the second-category text set are arranged from large to small according to the second similarity, the target similarity can refer to the similarity between the first second text and the first text; the seventh threshold can be set to 0.5.

[0121] Step 207 : For each second-category text set, when the target similarity in the second similarity satisfies the eighth threshold, the information processing device determines the second text and the first text in the second-category text set that meet the eighth threshold as the second optimized text set.

[0122] The optimized text set includes a first optimized text set and a second optimized text set.

[0123] Among them, the target similarity in the second similarity satisfies the eighth threshold, which may mean that at least one similarity in the target similarities in the second similarity is greater than the first sub-threshold, or multiple similarities in the target similarities in the second similarity are greater than the second sub-threshold; there are multiple texts in the second optimized text set.

[0124] In the embodiment of the present application, when at least one of the target similarities in the second similarity is greater than the first sub-threshold, or multiple similarities in the target similarities in the second similarity are greater than the second sub-threshold, it indicates that there are clusterable texts in the second type of text set, and a breadth-first strategy can be used for iteration to determine the second text in the second type of text set that meets the eighth threshold and the first text in the second type of text set as a cluster (i.e., the second optimized text set); it should be noted that each second optimized text set can be represented by X={x1, x2, ..., x n}, where x1 can represent the vector corresponding to the first text in the second optimized text set, x2~x n It can represent the vector corresponding to the second text in the second optimized text set, and each vector is a l-dimensional vector; in a feasible implementation, the first sub-threshold can refer to 0.9, and the second sub-threshold can refer to 0.85.

[0125] Step 208: For each second-category text set, if the target similarity in the second similarity of the information processing device does not meet the seventh threshold and the eighth threshold, determine the fourth similarity between each to-be-processed text in the second-category text set that does not meet the seventh threshold and the eighth threshold and the text in the second optimized text set.

[0126] In an embodiment of the present application, the text to be processed may refer to the text in each second-category text set that neither satisfies the seventh threshold nor the eighth threshold; the fourth similarity may refer to the similarity between the text to be processed and the text in the second optimized text set; when the target similarity in the second similarity neither satisfies the seventh threshold nor the eighth threshold, the similarity between each text to be processed and each text in the second optimized text set can be determined.

[0127] Step 209: The information processing device determines the first target text in each second optimized text set.

[0128] The second text includes the text to be processed.

[0129] In an embodiment of the present application, the first target text may refer to a text determined from the second optimized text set, and the first target text may be determined from the second optimized text set based on a screening condition; it should be noted that each second optimized text set contains a first target text.

[0130] It should be noted that step 209 can be implemented in the following ways:

[0131] Step 209b1: The information processing device determines the density parameter and the distance parameter of the first text and each second text in the second optimized text set.

[0132] In the embodiment of the present application, X={x1, x2, ..., x n} to represent the second optimized text set for explanation, when determining any text x in the second optimized text set i When the density parameter and distance parameter are used, the Euclidean distance formula can be used first. Calculate x i and any text x in the second optimized text set j The Euclidean distance between them, where p represents an element in the x vector, and p = 1 represents the first element in the x vector; then, we can determine the element with x i The number of texts K whose similarity is greater than 0.9, and then determine the sum of the Euclidean distances D between the K texts, and then we can use the formula p = e D / K Calculate x i The density parameter of the corresponding text; then determine xi The distance parameter δ between the text with a higher density parameter i .

[0133] It should be noted that for x i , the closest distance to the text with a higher density parameter than it can be expressed as When x i When the text with the largest density parameter is the text in the second optimized text set, the distance parameter

[0134] Step 209b2: The information processing device determines the first target text based on the density parameter and the distance parameter.

[0135] In an embodiment of the present application, the text with the largest density parameter and the largest distance parameter can be determined from multiple density parameters and distance parameters corresponding to all texts in the second optimized text set as the first target text; it should be noted that the first target text is the central text in each second optimized text.

[0136] Step 210: The information processing device divides the text to be processed into a first optimized text set and a second optimized text set based on the fourth similarity and the first target text.

[0137] It should be noted that step 210 can be implemented in the following ways:

[0138] In the embodiment of the present application, the text to be processed may be divided based on the fourth similarity first, and then divided based on the first target text.

[0139] Step 210c1: When there are P fourth similarities among the fourth similarities that are greater than a ninth threshold, the information processing device divides the text to be processed into a second optimized text set.

[0140] In a feasible implementation, when the total number of texts in the second optimized text set is T, P can be T / 2 (i.e., the number of half of the texts in the second optimized text); the ninth threshold can be set to 0.75; that is, when there are T / 2 fourth similarities greater than 0.75 in the fourth similarities, the text to be processed is divided into the second optimized text set.

[0141] Step 210c2: When there are no P fourth similarities greater than a ninth threshold value among the fourth similarities, the information processing device divides the text to be processed into a first optimized text set and a second optimized text set based on the first target text.

[0142] In a feasible implementation, if the fourth similarity of the texts in the second optimized text set does not exceed 0.75 for half of the texts, it is necessary to divide the text to be processed into the first optimized text set and the second optimized text set through the first target text.

[0143] It should be noted that the above embodiment may further include the following steps:

[0144] Step 211: The information processing device determines a total first occurrence number of texts in the first type of text set and a total second occurrence number of texts in the optimized text set.

[0145] In an embodiment of the present application, the first total number of occurrences refers to the frequency of occurrence of text in the first type of text set, and the second total number of occurrences refers to the frequency of occurrence of text in the second optimized text set; the frequency of occurrence of text in the first type of text set can be counted to obtain the total number of occurrences of text in the first type of text set, and the frequency of occurrence of text in the second optimized text set can be counted to obtain the total number of occurrences of text in the second optimized text set.

[0146] Step 212: The information processing device marks the first total number of occurrences in the first type of text set, and marks the second total number of occurrences in the optimized text set to obtain an updated text set.

[0147] In the embodiment of the present application, the first total number of occurrences is marked in the first type of text set, and the second total number of occurrences is marked in the optimized text set, and the following can be obtained: Figure 4 The text set (i.e., question cluster knowledge tree) with weight (i.e., frequency) is shown.

[0148] Step 213: The information processing device determines, for a third text in the updated text set whose total number of occurrences is less than the first number threshold, a fourth text at the same level as the third text whose total number of occurrences meets the second number threshold.

[0149] Among them, the updated text set includes the third text and the fourth text.

[0150] In an embodiment of the present application, the third text in the updated text set whose second total number of occurrences does not meet the first number threshold may refer to the text in the updated text set whose second total number of occurrences is less than the first number threshold, and the fourth text in which the second total number of occurrences meets the second number threshold may refer to the text in the updated text set whose second total number of occurrences is greater than the second number threshold; it should be noted that the first number threshold and the second number threshold may be determined based on historical experimental data; in a feasible implementation method, the first number threshold may be set to 20, and the second number threshold may be set to 100; the third text may refer to Figure 3 As shown in {What snacks can diabetics eat; What snacks can diabetics eat; What snacks are suitable for diabetics; ...}, and the total number of occurrences is 19, the fourth text may refer to Figure 3 The results are shown as {What fruits can people with diabetes eat?; What fruits are suitable for people with diabetes?; What fruits can people with diabetes eat; ...}, and the total number of occurrences is 116.

[0151] Step 214: The information processing device obtains a fifth text based on the second keyword in the third text and the characters other than the third keyword in the fourth text.

[0152] In an embodiment of the present application, the second keyword may refer to a keyword in a third text, and the third keyword may refer to a keyword in a fourth text; the keyword in the third text and the third keyword in the fourth text may be determined first, and then the third keyword may be removed from the fourth text, and the second keyword may be combined with characters other than the third keyword in the fourth text to obtain a fifth text.

[0153] In a feasible implementation, the third text is as follows Figure 3 As shown in {What snacks can diabetics eat; What snacks can diabetics eat; What snacks are suitable for diabetics; ...}, the fourth text is as follows Figure 3 In the case of {What fruits can diabetics eat?; What fruits are suitable for diabetics?; What fruits can diabetics eat; ...}, the second keyword may refer to snacks, and the third keyword may refer to fruits, then the fifth text {What snacks can diabetics eat?; What snacks are suitable for diabetics?; What snacks can diabetics eat; ...} can be obtained, and the total number of occurrences is 116, so the frequency of occurrence of the text where "snacks" is located can be greatly expanded.

[0154] It should be noted that the above embodiment may further include the following steps:

[0155] Step 215: The information processing device receives the query text input by the user.

[0156] In the embodiment of the present application, the inquiry text refers to the text input by the user; the inquiry text input by the user can be received in real time; in a feasible implementation method, for medical questions, grassroots patients (ie, users) can Figure 4 Enter the query text into the intelligent question and answer system in the digital platform shown, such as "For diabetic patients, what is the reason for only 4.6 uric acid levels?"

[0157] Step 216: The information processing device obtains a second target text from the target text set whose matching degree with the query text satisfies a first matching degree threshold, and displays response information of the second target text.

[0158] In an embodiment of the present application, the second target text may refer to a text in the target text set whose matching degree with the query text satisfies a first matching degree threshold; the matching degree of the query text satisfying the first matching degree threshold may refer to a text whose matching degree with the query text is greater than or equal to the first matching degree threshold; the response information may refer to the answer corresponding to the query text; the matching degree between each text in the target text set and the query text may be calculated layer by layer starting from the top layer according to the hierarchical relationship, and the first text in the target text set that is greater than the first matching degree threshold may be determined as the second target text, and the response information corresponding to the second target text may be displayed as the response information corresponding to the query text; it should be noted that the first matching degree threshold may be set based on historical experimental data.

[0159] In a feasible implementation, the first matching degree threshold may be set to 0.9; Figure 4 In the intelligent question-and-answer system in the digital platform shown, after entering the query text "Diabetic patients, what is the reason for uric acid levels of only 4.6?" and clicking the query button, the answer information "Diabetes and uric acid have no direct connection. Uric acid is a product of human purine metabolism. Under normal circumstances, the body's uric acid production and excretion are balanced. If the body produces too much and has no time to excrete it or the uric acid excretion mechanism degenerates, too much uric acid will be retained in the body. When the blood uric acid concentration is greater than 7 mg / dL, the body's body fluids become acidic, affecting the normal function of human cells. Long-term neglect will cause gout. In addition, excessive fatigue or lack of rest can also lead to relatively slow metabolism and cause gout. You need to pay attention to your diet, eat less animal offal (liver, kidney, bone marrow) and sardines, crabs, shrimps, spinach, large intestine, eat more vegetables, kelp, etc., eat a low-fat diet, drink less white wine, avoid beer, and drink more water." The digital platform provides an intelligent question-and-answer system that answers common medical consultation questions for grassroots patients. The system can effectively alleviate the doctor-patient ratio at the grassroots level and improve patients' medical experience.

[0160] Step 217: If there is no second target text in the target text set, the information processing device determines a third target text from the target text set whose matching degree with the query text satisfies a second matching degree threshold, and displays response information of the third target text.

[0161] In the embodiment of the present application, the third target text is a text in the target text set whose matching degree with the query text satisfies a second matching degree threshold; it should be noted that the second matching degree threshold may be determined based on historical experimental data;

[0162] In a feasible implementation, when the matching degree between the text in the target text set and the query text is less than the first matching degree threshold, it means that there is no second target text in the target text set. Then, multiple texts can be selected from the target text set according to the matching degree, and the multiple texts can be matched layer by layer in the order of the matching degree from large to small until the last layer is traversed. If the matching degree between the third target text and the query text is greater than or equal to the second matching degree threshold during the traversal process, the matching is terminated and the response information corresponding to the third target text is directly displayed.

[0163] The information processing method provided in the embodiment of the present application first determines a first type of text set having only one text to be analyzed and a second type of text set having multiple texts to be analyzed based on text features, and then only processes the second type of text set having multiple texts to be analyzed, thereby reducing the amount of data to be processed. When the second type of text set is processed again, multiple semantic similarity models are used to determine the similarity between texts, thereby improving the accuracy of determining the similarity between texts. Subsequently, a target text set with a hierarchical relationship is constructed, thereby improving the efficiency of retrieving user input questions in the target text set.

[0164] Based on the above embodiments, the embodiments of the present application provide an information processing method, referring to Figure 5 As shown, the method comprises the following steps:

[0165] Step 301: The information processing device determines the number of identical characters between the texts to be analyzed.

[0166] Step 302: The information processing device determines the difference between the total number of characters and the number of identical characters between the texts to be analyzed.

[0167] Step 303: The information processing device calculates the difference and the number of identical characters to obtain a first similarity.

[0168] It should be noted that after step 303, steps 304 to 306 may be executed, step 307 may be executed, steps 308 to 310 may be executed, or steps 311 to 312 may be executed.

[0169] Step 304: When the maximum similarity among the first similarities satisfies the first threshold and the first keyword does not exist in the text to be analyzed, the information processing device divides the text to be analyzed into a first category of text sets.

[0170] In an embodiment of the present application, when the first keyword does not exist in the text to be analyzed, it is directly determined that there is no problem cluster in the text to be analyzed, and each text to be analyzed is divided into an independent first-category text set; that is, the text to be analyzed is not clustered and can be clustered individually.

[0171] Step 305: When the maximum similarity among the first similarities meets the first threshold and the first keyword exists in the text to be analyzed, the information processing device removes the first keyword from the text to be analyzed and determines a third similarity between the texts after the keyword is deleted.

[0172] In the embodiment of the present application, the third similarity may refer to the similarity between texts after deleting the keyword; the similarity is calculated again after removing the first keyword from the text to be analyzed.

[0173] Step 306: When the maximum similarity among the third similarities satisfies the first threshold, the information processing device divides the text to be analyzed corresponding to the text after the keyword is deleted into a first type of text set.

[0174] In an embodiment of the present application, if the maximum similarity between the texts after deleting the keywords is still less than the first threshold, it is directly determined that there is no problem cluster in the text to be analyzed, and the text to be analyzed corresponding to the text after deleting the keywords is divided into an independent first-category text set.

[0175] Step 307: When the maximum similarity among the first similarities satisfies the second threshold, the information processing device determines that the text to be analyzed that satisfies the third threshold is the first text set.

[0176] Step 308: When the maximum similarity among the first similarities satisfies the fourth threshold and the number of the texts to be analyzed that satisfy the fifth threshold is greater than or equal to the first number, the information processing device divides the texts to be analyzed that satisfy the fifth threshold into a second text set.

[0177] In an embodiment of the present application, the number of texts to be analyzed that meet the fifth threshold can be counted. When the number of texts to be analyzed that meet the fifth threshold is greater than the first number, or the number of texts to be analyzed that meet the fifth threshold is equal to the first number, all texts to be analyzed that meet the fifth threshold are directly grouped as the first text set.

[0178] In a feasible implementation, the first number may be set to 1000.

[0179] Step 309: When the maximum similarity among the first similarities satisfies the fourth threshold and the number of the texts to be analyzed that satisfy the fifth threshold is less than the first number, the information processing device sorts the texts to be analyzed based on the first similarity to obtain second sorted texts.

[0180] In an embodiment of the present application, the second sorted text may refer to the text obtained after sorting the text to be analyzed based on the first similarity; when the number of texts to be analyzed that meet the fifth threshold is less than the first number, the texts to be analyzed can be sorted in descending order according to the first similarity to obtain the sorted text.

[0181] Step 310: The information processing device divides the first number of to-be-analyzed texts in the second sorted texts into a second text set.

[0182] In an embodiment of the present application, the first first number of texts to be analyzed can be obtained from the sorted texts, and the first first number of texts to be analyzed can be divided into a second text set, so that the number of texts to be analyzed in the second text set can be increased.

[0183] Step 311: When the maximum similarity among the first similarities satisfies a sixth threshold, the information processing device sorts the texts to be analyzed based on the first similarities to obtain first sorted texts.

[0184] Step 312: The information processing device determines the M texts to be analyzed in the first sorted texts as a third text set.

[0185] Among them, the second type of text collection includes the first text collection, the second text collection and the third text collection.

[0186] Step 313: The information processing device uses multiple semantic similarity models to determine the second similarity between the first text and each second text in each second type of text set.

[0187] Each second-category text set includes a first text and a plurality of second texts.

[0188] In a feasible implementation, when the first text is “At what age is diabetes most likely to occur?”, the multiple second texts are “At what age is diabetes most likely to occur?”, “At what age does diabetes develop?”, ... and “What kind of wine can diabetics drink during the Chinese New Year?”, and the two semantic similarity models used are ESIM and Linkage, the second similarity between “At what age is diabetes most likely to occur?” and “At what age is diabetes most likely to occur?” determined by ESIM and Linkage is (0.87, 0.611), the second similarity between “At what age is diabetes most likely to occur?” and “At what age does diabetes develop?” is (0.826, 0.7), and the second similarity between “At what age is diabetes most likely to occur?” and “What kind of wine can diabetics drink during the Chinese New Year?” is (0.25, 0.5); When the first text is "Will being nervous affect blood sugar?", the multiple second texts are "Will always being nervous cause diabetes?", "Will emotional tension cause increased blood sugar?", ... and "Will glucose affect calcium absorption?", and the two semantic similarity models used are ESIM and Linkage, the second similarity between "Will being nervous affect blood sugar" and "Will always being nervous cause diabetes?" determined by ESIM and Linkage is (0.856, 0.71), the second similarity between "Will being nervous affect blood sugar" and "Will emotional tension cause increased blood sugar?" is (0.851, 0.892), and the second similarity between "Will being nervous affect blood sugar" and "Will glucose affect calcium absorption?" is (0.128, 0.242).

[0189] Step 314: For each second-category text set, the information processing device determines that the first text and each second text in the second-category text set are each a first optimized text set when the target similarity in the second similarity satisfies a seventh threshold.

[0190] In a feasible implementation, the first text is "I have been feeling sleepy and cold after taking fructose. What is going on?", and the multiple second texts are "What is fructose?", "Why do I have diarrhea after taking lactulose?", ... and "What fruits can and cannot diabetics eat?", and the two semantic similarity models used are ESIM and Linkage, then the second similarity between "I have been feeling sleepy and cold after taking fructose. What is going on?" and "What is fructose?" determined by ESIM and Linkage is (0.346, 0.367), and the second similarity between "I have been feeling sleepy and cold after taking fructose. What is going on?" and "Why do I have diarrhea after taking lactulose?" is (0.350, 0.369). The second similarity is (0.25, 0.335), and the second similarity between "I feel sleepy and cold after taking fructose. What's going on" and "What fruits can and cannot diabetics eat?" is (0.0006, 0.167); the target similarities (0.346, 0.367) are both less than the seventh threshold of 0.5, so it is determined that "I feel sleepy and cold after taking fructose. What's going on" is a first optimized text set, "What is fructose?" is a first optimized text set, "Why do I have diarrhea after taking lactulose?" is a first optimized text set, and..., "What fruits can and cannot diabetics eat?" is a first optimized text set.

[0191] Step 315: For each second-category text set, when the target similarity in the second similarity satisfies the eighth threshold, the information processing device determines the second text and the first text in the second-category text set that meet the eighth threshold as the second optimized text set.

[0192] The optimized text set includes a first optimized text set and a second optimized text set.

[0193] In a feasible implementation, the first text is "What does excessive insulin cause?", the multiple second texts are "What will cause excessive insulin", "What happens if insulin is too high", ... and "What is glucagonoma?", and the two semantic similarity models used are ESIM and Linkage, then the second similarity between "What does excessive insulin cause?" and "What will cause excessive insulin" determined by ESIM and Linkage is (0.998, 0.964), the second similarity between "What does excessive insulin cause?" and "What happens if insulin is too high" is (0.644, 0.766), and the second similarity between "What does excessive insulin cause?" and "What is glucagonoma?" is ((0.028, 0.35); then the target similarity (0.998, 0.964) meets the eighth threshold, and it is determined that "What does excessive insulin cause?", "What does excessive insulin cause?", "What happens if insulin is too high", ... and "What is glucagonoma?" are the second optimized text set.

[0194] Step 316: For each second-category text set, if the target similarity in the second similarity of the information processing device does not meet the seventh threshold and the eighth threshold, determine the fourth similarity between each to-be-processed text in the second-category text set that does not meet the seventh threshold and the eighth threshold and the text in the second optimized text set.

[0195] In a feasible implementation, the first text is "What does excessive insulin cause?", the multiple second texts are "What will cause excessive insulin", "What happens if insulin is too high", ... and "What is glucagonoma?", and the two semantic similarity models used are ESIM and Linkage, then the second similarity between "What does excessive insulin cause?" and "What will cause excessive insulin" determined by ESIM and Linkage is (0.998, 0.964), the second similarity between "What does excessive insulin cause?" and "What happens if insulin is too high" is (0.644, 0.766), and the second similarity between "What does excessive insulin cause?" and "What is glucagonoma?" is ((0.028, 0.35); then the target similarity (0.998, 0.964) meets the eighth threshold, and it is determined that "What does excessive insulin cause?", "What does excessive insulin cause?", "What happens if insulin is too high", ... and "What is glucagonoma?" are the second optimized text set.

[0196] Step 317: The information processing device determines a density parameter and a distance parameter of the first text and each second text in the second optimized text set.

[0197] Step 318: The information processing device determines the first target text based on the density parameter and the distance parameter.

[0198] Step 319: When there are P fourth similarities among the fourth similarities that are greater than a ninth threshold, the information processing device divides the text to be processed into a second optimized text set.

[0199] Step 320: When there are no P fourth similarities greater than a ninth threshold value among the fourth similarities, the information processing device divides the text to be processed into a first optimized text set and a second optimized text set based on the first target text.

[0200] It should be noted that step 320 can be implemented in the following ways:

[0201] Step 320d1: The information processing device determines the fifth similarity between each to-be-processed text and each first target text.

[0202] In the embodiment of the present application, the fifth similarity may refer to the similarity between the text to be processed and the first target text; a pre-trained model may be used to calculate the similarity between each text to be processed and each first target text.

[0203] Step 320d2: When the maximum similarity among the fifth similarities is less than or equal to the second similarity corresponding to the first target text, the information processing device divides the text to be processed into a second optimized text set.

[0204] In an embodiment of the present application, the second similarity corresponding to the first target text may refer to the maximum similarity boundary corresponding to the second optimized text set; when the maximum similarity in the fifth similarity is less than or equal to the second similarity corresponding to the first target text, the text to be processed is directly added to the second optimized text set.

[0205] Step 320d3: When the maximum similarity among the fifth similarities is greater than the second similarity corresponding to the first target text, the information processing device divides the text to be processed into a first optimized text set.

[0206] In an embodiment of the present application, when the maximum similarity among the fifth similarities is greater than the second similarity corresponding to the first target text, the text to be processed can be directly added to the first optimized text set, or the text to be processed can be divided based on experience to determine whether there is any text in the text to be processed that can be divided into the second optimized text set. If not, the text to be processed can be added to the first optimized text set.

[0207] In other embodiments of the present application, due to the particularity of the text in the medical field (i.e., medical questions), a more detailed knowledge tree with a hierarchical relationship (i.e., a target text set) can be constructed to represent the content of each question cluster. For example, diabetes can be clinically divided into type 1 diabetes, type 2 diabetes, special types of diabetes, and gestational diabetes, and there are multiple question clusters in practice, such as "How to treat diabetes", "How to treat type 1 diabetes", etc. By constructing a knowledge tree with a hierarchical relationship, on the one hand, the answers to medical questions can be effectively managed in a refined manner.

[0208] The following six categories of questions (i.e., texts) have the highest frequency of occurrence, namely (1) prevention, treatment, inheritance, and hazards of different types of diabetes. The parent category includes how to prevent and treat diabetes, what are the symptoms of diabetes, etc., and the subcategory is the specific classification type; (2) treatment methods for abnormal blood sugar values, including high blood sugar and low blood sugar, and then the specific blood sugar values ​​are divided into high and low according to the diabetes diagnostic criteria; (3) diet-related issues. The parent category includes fruits, staple foods, vegetables, etc., and then according to the specific situation, fruits include watermelon, peach, etc.; staple foods include rice, noodles, etc.; vegetables include fungus, celery, etc.; (4) the usage and contraindications of common drugs, such as metformin, gliclazide, and glycopyrrolate; (5) consultation on issues related to diabetes complications, including diabetic foot, diabetic retinopathy, diabetic nephropathy, etc.; (6) consultation on whether the blood sugar level in a certain age group is normal, such as "What is the standard blood sugar level for a 60-year-old?", "What is the normal blood sugar level for a 65-year-old?", "What is the normal blood sugar level for a 70-year-old?", etc.

[0209] In order to construct the most frequent questions about the above six categories, we can first build a question classification model to divide the questions into the above six major subcategories + other categories. For each, we select the central cluster sample as the representative, and then use certain rules to map the inter-cluster questions to the corresponding subtrees, and get the following Figure 3 The diabetes question cluster knowledge tree is shown.

[0210] The information processing method provided in the embodiment of the present application first determines a first type of text set having only one text to be analyzed and a second type of text set having multiple texts to be analyzed based on text features, and then only processes the second type of text set having multiple texts to be analyzed, thereby reducing the amount of data to be processed. When the second type of text set is processed again, multiple semantic similarity models are used to determine the similarity between texts, thereby improving the accuracy of determining the similarity between texts. Subsequently, a target text set with a hierarchical relationship is constructed, thereby improving the efficiency of retrieving user input questions in the target text set.

[0211] Based on the above embodiments, the embodiments of the present application provide an information processing device, which can be applied to Figure 1 , Figure 2 and Figure 5 In the information processing method provided in the corresponding embodiment, refer to Figure 6 As shown, the device 4 may include: a first processing unit 41, a determining unit 42, a second processing unit 43 and a generating unit 44, wherein:

[0212] A first processing unit 41 is used to determine a first similarity between the texts to be analyzed based on the text features of the texts to be analyzed, and divide the multiple texts to be analyzed into multiple text sets based on the first similarity; wherein the multiple text sets include a first type of text set having one text to be analyzed and a second type of text set having multiple texts to be analyzed;

[0213] A determination unit 42 is used to determine the second similarity between the first text and each second text in each second-category text set by using multiple semantic similarity models; wherein each second-category text set includes the first text and multiple second texts;

[0214] A second processing unit 43 is used to process the plurality of second texts in each second type text set based on the second similarity to obtain an optimized text set;

[0215] The generating unit 44 is used to generate a target text set with a hierarchical relationship for the text to be analyzed based on the first type of text set, the optimized text set and the target hierarchical information.

[0216] In the embodiment of the present application, the first processing unit 41 is further configured to perform the following steps:

[0217] Determine the number of identical characters between the texts to be analyzed;

[0218] Determine the difference between the total number of characters and the number of identical characters between the texts to be analyzed;

[0219] Calculate the difference and the number of identical characters to obtain a first similarity;

[0220] Based on the first similarity and the similarity threshold, the plurality of texts to be analyzed are divided into a plurality of text sets.

[0221] In the embodiment of the present application, the first processing unit 41 is further configured to perform the following steps:

[0222] When the maximum similarity among the first similarities meets the first threshold, dividing the plurality of texts to be analyzed into a first category of text sets based on the first keyword;

[0223] When the maximum similarity satisfies the second threshold, determining the text to be analyzed that satisfies the third threshold as the first text set;

[0224] When the maximum similarity satisfies the fourth threshold, the texts to be analyzed are divided into a second text set based on the number of texts to be analyzed that meet the fifth threshold;

[0225] When the maximum similarity satisfies the sixth threshold, sorting the texts to be analyzed based on the first similarity to obtain first sorted texts;

[0226] The M texts to be analyzed in the first sorted texts are determined as a third text set; wherein the second type of text set includes the first text set, the second text set and the third text set.

[0227] In the embodiment of the present application, the first processing unit 41 is further configured to perform the following steps:

[0228] In the case that the first keyword does not exist in the text to be analyzed, the text to be analyzed is divided into a first category of text sets;

[0229] In the case where the first keyword exists in the text to be analyzed, deleting the first keyword in the text to be analyzed, and determining a third similarity between the texts after the keyword is deleted;

[0230] When the maximum similarity among the third similarities meets the first threshold, the text to be analyzed corresponding to the text after the keyword is deleted is divided into a first type of text set.

[0231] In the embodiment of the present application, the first processing unit 41 is further configured to perform the following steps:

[0232] When the number of the to-be-analyzed texts satisfying the fifth threshold is greater than or equal to the first number, dividing the to-be-analyzed texts satisfying the fifth threshold into a second text set;

[0233] When the number of the to-be-analyzed texts satisfying the fifth threshold is less than the first number, sorting the to-be-analyzed texts based on the first similarity to obtain second sorted texts;

[0234] A first number of to-be-analyzed texts in the second sorted texts are divided into a second text set.

[0235] In the embodiment of the present application, the second processing unit 43 is further configured to perform the following steps:

[0236] For each second-category text set, when the target similarity in the second similarity satisfies a seventh threshold, determining that the first text and each second text in the second-category text set are each a first optimized text set;

[0237] For each second-category text set, when the target similarity in the second similarity satisfies the eighth threshold, the second text and the first text in the second-category text set that meet the eighth threshold are determined as the second optimized text set; wherein the optimized text set includes the first optimized text set and the second optimized text set.

[0238] In the embodiment of the present application, the second processing unit 43 is further configured to perform the following steps:

[0239] For each second-category text set, if the target similarity in the second similarity does not meet the seventh threshold and the eighth threshold, determine the fourth similarity between each to-be-processed text in the second-category text set that does not meet the seventh threshold and the eighth threshold and the text in the second optimized text set;

[0240] Determine a first target text in each second optimized text set;

[0241] Based on the fourth similarity and the first target text, the text to be processed is divided into a first optimized text set and a second optimized text set.

[0242] In the embodiment of the present application, the second processing unit 43 is further configured to perform the following steps:

[0243] Determining a density parameter and a distance parameter of the first text and each second text in the second optimized text set;

[0244] A first target text is determined based on the density parameter and the distance parameter.

[0245] In the embodiment of the present application, the second processing unit 43 is further configured to perform the following steps:

[0246] In the case that there are P fourth similarities among the fourth similarities that are greater than the ninth threshold, dividing the text to be processed into a second optimized text set;

[0247] When there are no P fourth similarities among the fourth similarities that are greater than the ninth threshold, the text to be processed is divided into a first optimized text set and a second optimized text set based on the first target text.

[0248] In the embodiment of the present application, the second processing unit 43 is further configured to perform the following steps:

[0249] Determining a fifth similarity between each to-be-processed text and each first target text;

[0250] When the maximum similarity among the fifth similarities is less than or equal to the second similarity corresponding to the first target text, dividing the text to be processed into a second optimized text set;

[0251] When the maximum similarity among the fifth similarities is greater than the second similarity corresponding to the first target text, the text to be processed is divided into a first optimized text set.

[0252] In the embodiment of the present application, the second processing unit 43 is further configured to perform the following steps:

[0253] Determine the total number of first occurrences of text in the first type of text set and the total number of second occurrences of text in the optimized text set;

[0254] The first total number of occurrences is marked in the first type of text set, and the second total number of occurrences is marked in the optimized text set to obtain an updated text set.

[0255] In the embodiment of the present application, the second processing unit 43 is further configured to perform the following steps:

[0256] For a third text in the updated text set whose total number of occurrences is less than the first number threshold, determining a fourth text at the same level as the third text whose total number of occurrences meets the second number threshold; wherein the updated text set includes the third text and the fourth text;

[0257] A fifth text is obtained based on the second keyword in the third text and characters other than the third keyword in the fourth text.

[0258] In the embodiment of the present application, the second processing unit 43 is further configured to perform the following steps:

[0259] Receive query text input by the user;

[0260] Acquire a second target text from the target text set, the matching degree of which satisfies a first matching degree threshold with the query text, and display the response information of the second target text;

[0261] In the case that the second target text does not exist in the target text set, a third target text whose matching degree with the query text satisfies a second matching degree threshold is determined from the target text set, and response information of the third target text is displayed.

[0262] It should be noted that the specific implementation process of the steps performed by each unit in this embodiment can be referred to Figure 1 , Figure 2 and Figure 5 The implementation process of the information processing method provided in the corresponding embodiment will not be repeated here.

[0263] The information processing device provided by the embodiment of the present application first determines a first type of text set with only one text to be analyzed and a second type of text set with multiple texts to be analyzed based on text features, and then only processes the second type of text set with multiple texts to be analyzed, thereby reducing the amount of processed data, and when reprocessing the second type of text set, multiple semantic similarity models are used to determine the similarity between texts, thereby improving the accuracy of determining the similarity between texts, and then constructs a target text set with a hierarchical relationship, thereby improving the efficiency of retrieving user input questions in the target text set.

[0264] Based on the above embodiments, the embodiments of the present application provide an information processing device, which can be applied to Figure 1 , Figure 2 and Figure 5 The information processing method provided by the corresponding embodiment refers to Figure 7 As shown, the device 5 may include a processor 51, a memory 52 and a communication bus 53;

[0265] The communication bus 53 is used to realize the communication connection between the processor 51 and the memory 52;

[0266] The processor 51 is used to execute the information processing program stored in the memory 52 to implement the following steps:

[0267] Determine a first similarity between the texts to be analyzed based on the text features of the texts to be analyzed, and divide the multiple texts to be analyzed into multiple text sets based on the first similarity; wherein the multiple text sets include a first type of text set having one text to be analyzed and a second type of text set having multiple texts to be analyzed;

[0268] Using multiple semantic similarity models, determine the second similarity between the first text and each second text in each second-category text set; wherein each second-category text set includes the first text and multiple second texts;

[0269] Processing the plurality of second texts in each second type of text set based on the second similarity to obtain an optimized text set;

[0270] Based on the first type of text set, the optimized text set and the target hierarchical information, a target text set with a hierarchical relationship for the text to be analyzed is generated.

[0271] In other embodiments of the present application, the processor 51 is used to execute the memory 52 to determine the first similarity between the texts to be analyzed based on the text features of the texts to be analyzed, and to divide the multiple texts to be analyzed into multiple text sets based on the first similarity, so as to implement the following steps:

[0272] Determine the number of identical characters between the texts to be analyzed;

[0273] Determine the difference between the total number of characters and the number of identical characters between the texts to be analyzed;

[0274] Calculate the difference and the number of identical characters to obtain a first similarity;

[0275] Based on the first similarity and the similarity threshold, the plurality of texts to be analyzed are divided into a plurality of text sets.

[0276] In other embodiments of the present application, the processor 51 is used to execute the memory 52 to divide the multiple texts to be analyzed into multiple text sets based on the first similarity and the similarity threshold, so as to implement the following steps:

[0277] When the maximum similarity among the first similarities meets the first threshold, dividing the plurality of texts to be analyzed into a first category of text sets based on the first keyword;

[0278] When the maximum similarity satisfies the second threshold, determining the text to be analyzed that satisfies the third threshold as the first text set;

[0279] When the maximum similarity satisfies the fourth threshold, the texts to be analyzed are divided into a second text set based on the number of texts to be analyzed that meet the fifth threshold;

[0280] When the maximum similarity satisfies the sixth threshold, sorting the texts to be analyzed based on the first similarity to obtain first sorted texts;

[0281] The M texts to be analyzed in the first sorted texts are determined as a third text set; wherein the second type of text set includes the first text set, the second text set and the third text set.

[0282] In other embodiments of the present application, the processor 51 is used to execute the division of the plurality of to-be-analyzed texts into a first category of text sets based on the first keyword in the memory 52 to implement the following steps:

[0283] In the case that the first keyword does not exist in the text to be analyzed, the text to be analyzed is divided into a first category of text sets;

[0284] In the case where the first keyword exists in the text to be analyzed, deleting the first keyword in the text to be analyzed, and determining a third similarity between the texts after the keyword is deleted;

[0285] When the maximum similarity among the third similarities meets the first threshold, the text to be analyzed corresponding to the text after the keyword is deleted is divided into a first type of text set.

[0286] In other embodiments of the present application, the processor 51 is used to execute in the memory 52, when the maximum similarity satisfies the fourth threshold, based on the number of texts to be analyzed that meet the fifth threshold, to divide the text to be analyzed into a second text set, so as to implement the following steps:

[0287] When the number of the to-be-analyzed texts satisfying the fifth threshold is greater than or equal to the first number, dividing the to-be-analyzed texts satisfying the fifth threshold into a second text set;

[0288] When the number of the to-be-analyzed texts satisfying the fifth threshold is less than the first number, sorting the to-be-analyzed texts based on the first similarity to obtain second sorted texts;

[0289] A first number of to-be-analyzed texts in the second sorted texts are divided into a second text set.

[0290] In other embodiments of the present application, the processor 51 is used to execute the memory 52 to process the multiple second texts in each second type text set based on the second similarity to obtain an optimized text set, so as to implement the following steps:

[0291] For each second-category text set, when the target similarity in the second similarity satisfies a seventh threshold, determining that the first text and each second text in the second-category text set are each a first optimized text set;

[0292] For each second-category text set, when the target similarity in the second similarity satisfies the eighth threshold, the second text and the first text in the second-category text set that meet the eighth threshold are determined as the second optimized text set; wherein the optimized text set includes the first optimized text set and the second optimized text set.

[0293] In other embodiments of the present application, the processor 51 is used to execute the information processing program stored in the memory 52 to implement the following steps:

[0294] For each second-category text set, if the target similarity in the second similarity does not meet the seventh threshold and the eighth threshold, determine the fourth similarity between each to-be-processed text in the second-category text set that does not meet the seventh threshold and the eighth threshold and the text in the second optimized text set;

[0295] Determine a first target text in each second optimized text set;

[0296] Based on the fourth similarity and the first target text, the text to be processed is divided into a first optimized text set and a second optimized text set.

[0297] In other embodiments of the present application, the processor 51 is used to execute the determination of the first target text in each second optimized text set in the memory 52 to implement the following steps:

[0298] Determining a density parameter and a distance parameter of the first text and each second text in the second optimized text set;

[0299] A first target text is determined based on the density parameter and the distance parameter.

[0300] In other embodiments of the present application, the processor 51 is used to execute in the memory 52, based on the fourth similarity and the first target text, to divide the text to be processed into a first optimized text set and a second optimized text set, so as to implement the following steps:

[0301] In the case that there are P fourth similarities among the fourth similarities that are greater than the ninth threshold, dividing the text to be processed into a second optimized text set;

[0302] When there are no P fourth similarities among the fourth similarities that are greater than the ninth threshold, the text to be processed is divided into a first optimized text set and a second optimized text set based on the first target text.

[0303] In other embodiments of the present application, the processor 51 is used to execute the memory 52 to divide the to-be-processed text into a first optimized text set and a second optimized text set based on the first target text, so as to implement the following steps:

[0304] Determining a fifth similarity between each to-be-processed text and each first target text;

[0305] When the maximum similarity among the fifth similarities is less than or equal to the second similarity corresponding to the first target text, dividing the text to be processed into a second optimized text set;

[0306] When the maximum similarity among the fifth similarities is greater than the second similarity corresponding to the first target text, the text to be processed is divided into a first optimized text set.

[0307] In other embodiments of the present application, the processor 51 is used to execute the information processing program stored in the memory 52 to implement the following steps:

[0308] Determine the total number of first occurrences of text in the first type of text set and the total number of second occurrences of text in the optimized text set;

[0309] The first total number of occurrences is marked in the first type of text set, and the second total number of occurrences is marked in the optimized text set to obtain an updated text set.

[0310] In other embodiments of the present application, the processor 51 is used to execute the information processing program stored in the memory 52 to implement the following steps:

[0311] For a third text in the updated text set whose total number of occurrences is less than the first number threshold, determining a fourth text at the same level as the third text whose total number of occurrences meets the second number threshold; wherein the updated text set includes the third text and the fourth text;

[0312] A fifth text is obtained based on the second keyword in the third text and characters other than the third keyword in the fourth text.

[0313] In other embodiments of the present application, the processor 51 is used to execute the information processing program stored in the memory 52 to implement the following steps:

[0314] Receive query text input by the user;

[0315] Acquire a second target text from the target text set, the matching degree of which satisfies a first matching degree threshold with the query text, and display the response information of the second target text;

[0316] In the case that the second target text does not exist in the target text set, a third target text whose matching degree with the query text satisfies a second matching degree threshold is determined from the target text set, and response information of the third target text is displayed.

[0317] It should be noted that the specific implementation process of the steps executed by the processor in this embodiment can refer to Figure 1 , Figure 2 and Figure 5 The implementation process of the information processing method provided in the corresponding embodiment will not be repeated here.

[0318] The information processing device provided by the embodiment of the present application first determines a first type of text set having only one text to be analyzed and a second type of text set having multiple texts to be analyzed based on text features, and then only processes the second type of text set having multiple texts to be analyzed, thereby reducing the amount of data to be processed. When the second type of text set is processed again, multiple semantic similarity models are used to determine the similarity between texts, thereby improving the accuracy of determining the similarity between texts. A target text set with a hierarchical relationship is then constructed, thereby improving the efficiency of retrieving user input questions in the target text set.

[0319] Based on the foregoing embodiments, the embodiments of the present application provide a computer-readable storage medium, which stores one or more programs, and the one or more programs can be executed by one or more processors to implement Figure 1 , Figure 2 and Figure 5The corresponding embodiments provide steps in the information processing method.

[0320] It should be noted that the above-mentioned computer-readable storage medium can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a magnetic random access memory (FRAM), a flash memory (Flash Memory), a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM) and other memories; it can also be various electronic devices including one or any combination of the above-mentioned memories, such as mobile phones, computers, tablet devices, personal digital assistants, etc.

[0321] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

[0322] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0323] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course, by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0324] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products of the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0325] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0326] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0327] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. An information processing method, characterized in that: The method comprises: Determine a first similarity between the texts to be analyzed based on the text features of the texts to be analyzed, and divide the plurality of texts to be analyzed into a plurality of text sets based on the first similarity; wherein the plurality of text sets include a first type of text set having one text to be analyzed and a second type of text set having a plurality of texts to be analyzed; Using multiple semantic similarity models, determine the second similarity between the first text and each second text in each second-category text set; wherein each second-category text set includes the first text and multiple second texts; Processing the plurality of second texts in each of the second-category text sets based on the second similarity to obtain an optimized text set; Based on the first type of text set, the optimized text set and the target hierarchical information, a target text set with a hierarchical relationship for the text to be analyzed is generated.

2. The method according to claim 1, characterized in that The step of determining a first similarity between the texts to be analyzed based on the text features of the texts to be analyzed, and dividing the plurality of texts to be analyzed into a plurality of text sets based on the first similarity, comprises: Determine the number of identical characters between the texts to be analyzed; Determine the difference between the total number of characters in the text to be analyzed and the number of the same characters; Calculating the difference and the number of identical characters to obtain the first similarity; Based on the first similarity and the similarity threshold, the multiple texts to be analyzed are divided into the multiple text sets.

3. The method according to claim 2, characterized in that The dividing the plurality of texts to be analyzed into the plurality of text sets based on the first similarity and the similarity threshold comprises: When the maximum similarity among the first similarities meets a first threshold, dividing the plurality of texts to be analyzed into the first category of text sets based on a first keyword; In the case where the maximum similarity satisfies the second threshold, determining the text to be analyzed that satisfies the third threshold as the first text set; In the case where the maximum similarity satisfies a fourth threshold, based on the number of texts to be analyzed that satisfy a fifth threshold, dividing the texts to be analyzed into a second text set; In a case where the maximum similarity satisfies a sixth threshold, sorting the text to be analyzed based on the first similarity to obtain a first sorted text; The M texts to be analyzed in the first sorted texts are determined as a third text set; wherein the second type of text set includes the first text set, the second text set and the third text set.

4. The method according to claim 3, characterized in that The step of dividing the plurality of texts to be analyzed into the first type of text set based on the first keyword includes: In the case that the first keyword does not exist in the text to be analyzed, dividing the text to be analyzed into a first type of text set; In the case where the first keyword exists in the text to be analyzed, deleting the first keyword in the text to be analyzed, and determining a third similarity between the texts after the keyword is deleted; When the maximum similarity among the third similarities meets the first threshold, the text to be analyzed corresponding to the text after the keyword is deleted is divided into a first type of text set.

5. The method according to claim 3, characterized in that: When the maximum similarity satisfies the fourth threshold, dividing the text to be analyzed into a second text set based on the number of texts to be analyzed that meet the fifth threshold comprises: When the number of the to-be-analyzed texts satisfying the fifth threshold is greater than or equal to the first number, dividing the to-be-analyzed texts satisfying the fifth threshold into the second text set; In a case where the number of the to-be-analyzed texts satisfying the fifth threshold is less than the first number, sorting the to-be-analyzed texts based on the first similarity to obtain second sorted texts; The first number of to-be-analyzed texts in the second sorted texts are divided into the second text set.

6. The method according to claim 1, characterized in that The step of processing the plurality of second texts in each second type text set based on the second similarity to obtain an optimized text set includes: For a second type of text set whose target similarity in the second similarity satisfies a seventh threshold, determining that the first text and each of the second texts in the second type of text set are each a first optimized text set; For a second type of text set whose target similarity in the second similarity meets an eighth threshold, a second text in the second type of text set that meets the eighth threshold and the first text are determined as a second optimized text set; wherein the optimized text set includes the first optimized text set and the second optimized text set.

7. The method according to claim 6, characterized in that The method further comprises: For the second type of text set whose target similarity in the second similarity does not satisfy the seventh threshold and the eighth threshold, determining a fourth similarity between each to-be-processed text in the second type of text set that does not satisfy the seventh threshold and the eighth threshold and a text in the second optimized text set; Determine a first target text in each of the second optimized text sets; Based on the fourth similarity and the first target text, the to-be-processed text is divided into the first optimized text set and the second optimized text set.

8. The method according to claim 7, characterized in that The determining of the first target text in each of the second optimized text sets comprises: Determining a density parameter and a distance parameter of the first text and each of the second texts in the second optimized text set; The first target text is determined based on the density parameter and the distance parameter.

9. The method according to claim 7, characterized in that: The dividing the to-be-processed text into the first optimized text set and the second optimized text set based on the fourth similarity and the first target text comprises: In the case where there are P fourth similarities among the fourth similarities that are greater than a ninth threshold, dividing the to-be-processed text into the second optimized text set; In the case that there are not P fourth similarities among the fourth similarities that are greater than the ninth threshold, the to-be-processed text is divided into the first optimized text set and the second optimized text set based on the first target text.

10. The method according to claim 9, characterized in that The step of dividing the to-be-processed text into the first optimized text set and the second optimized text set based on the first target text includes: Determining a fifth similarity between each of the to-be-processed texts and each of the first target texts; When the maximum similarity among the fifth similarities is less than or equal to the second similarity corresponding to the first target text, dividing the to-be-processed text into the second optimized text set; When the maximum similarity among the fifth similarities is greater than the second similarity corresponding to the first target text, the text to be processed is divided into the first optimized text set.

11. The method according to claim 1, characterized in that: The method further comprises: Determine a total number of first occurrences of text in the first type of text set and a total number of second occurrences of text in the optimized text set; The first total number of occurrences is marked in the first type of text set, and the second total number of occurrences is marked in the optimized text set to obtain an updated text set.

12. The method according to claim 11, characterized in that The method further comprises: For a third text in the updated text set whose total number of occurrences is less than the first number threshold, determine a fourth text at the same level as the third text whose total number of occurrences meets the second number threshold; wherein the updated text set includes the third text and the fourth text; A fifth text is obtained based on the second keyword in the third text and the characters other than the third keyword in the fourth text.

13. The method according to claim 1, characterized in that The method further comprises: Receive query text input by the user; Acquire a second target text from the target text set, the matching degree of which satisfies a first matching degree threshold with the query text, and display response information of the second target text; In the case that the second target text does not exist in the target text set, a third target text whose matching degree with the query text satisfies a second matching degree threshold is determined from the target text set, and response information of the third target text is displayed.

14. An information processing device, characterized in that: The device comprises: A first processing unit is used to determine a first similarity between the texts to be analyzed based on text features of the texts to be analyzed, and divide the plurality of texts to be analyzed into a plurality of text sets based on the first similarity; wherein the plurality of text sets include a first type of text set having one text to be analyzed and a second type of text set having a plurality of texts to be analyzed; A determination unit, configured to determine a second similarity between the first text and each second text in each of the second-category text sets by using a plurality of semantic similarity models; wherein each of the second-category text sets includes the first text and a plurality of the second texts; A second processing unit, configured to process the plurality of second texts in each of the second-category text sets based on the second similarity to obtain an optimized text set; A generating unit is used to generate a target text set with a hierarchical relationship for the text to be analyzed based on the first type of text set, the optimized text set and the target hierarchical information.

15. An information processing device, characterized in that: The device comprises: a processor, a memory and a communication bus; The communication bus is used to realize the communication connection between the processor and the memory; The processor is used to execute the information processing program stored in the memory to implement the steps of the information processing method according to any one of claims 1 to 13.

16. A computer-readable storage medium, characterized in that: The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of the information processing method according to any one of claims 1 to 13.

Images