Project judgment method, related equipment and readable storage medium

A judging method and project technology, applied in the direction of neural learning methods, instruments, biological neural network models, etc., can solve the problems of large manpower and time, low efficiency, high cost, etc., and achieve the effects of enhancing transparency, reducing costs, and strong objectivity

Pending Publication Date: 2020-05-12
TSINGHUA UNIV +1
0 Cites 2 Cited by

AI-Extracted Technical Summary

Problems solved by technology

[0004] However, manual judgment will consume a lot of...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Method used

The application discloses a kind of project judging method, at first obtain the project associated text of project to be judged, then determine the main information in the project associated text, and, the modification information of each main information; Based on the main information, and, every Finally, analyze the graph structure of the project-related text to determine the judgment result of the project to be judged. In this application, the ...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Abstract

The invention discloses a project judgment method, and the method comprises the steps: obtaining a project associated text of a to-be-judged project, and then determining trunk information in the project associated text and modification information of each piece of trunk information; determining a graph structure of the project associated text based on the trunk information and the modification information of each piece of trunk information; and finally, analyzing the graph structure of the project associated text, and determining a judgment result of the to-be-judged project. Compared with amanual judgment mode, the mode of realizing project judgment by analyzing the graph structure of the project associated text can save manpower and time, so the cost can be reduced, and the project judgment efficiency can be improved. Based on the above mode, the judgment of items such as auxiliary diagnosis prediction items in the medical field and auxiliary verification value prediction items inthe judicial field can be realized.

Application Domain

Technology Topic

Image

  • Project judgment method, related equipment and readable storage medium
  • Project judgment method, related equipment and readable storage medium
  • Project judgment method, related equipment and readable storage medium

Examples

  • Experimental program(1)

Example Embodiment

[0071] The technical solutions in the embodiments of the present application will be described clearly and completely in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of this application.
[0072] Next, the project determination method provided in this application is introduced through the following embodiments.
[0073] See figure 1 , figure 1 This is a schematic flow diagram of the project determination method disclosed in the application embodiments, and the method may include:
[0074] S101: Obtain the item-related text of the item to be determined.
[0075] In this application, the items to be determined may be items in various fields, and specifically may be auxiliary prediction items, such as auxiliary diagnosis prediction items in the medical field, and auxiliary verification and value prediction items in the judicial field. Different projects have different project-related texts. For example, the project-related text of the auxiliary diagnosis and prediction project in the medical field can be electronic medical record text, and the project-related text of the auxiliary verification and value prediction project in the judicial field can be the evidence materials of judicial cases.
[0076] It should be noted that in this application, the item-related text of the item to be determined is a text that meets the text format specified by the item determination method, and can be the original item-related text of the item to be determined. If the original item-related text of the item to be determined is not To meet the text format specified in the item determination method, the original item-related text needs to be preprocessed, and the pre-processed item-related text is used as the item-related text of the item to be determined.
[0077] As an implementable manner, in this application, the text format that meets the requirements of the project specification method can be preset to be the json format.
[0078] S102: Determine the main information in the related text of the item and the modification information of each main information.
[0079] In this application, the item-related text of the item to be determined is different from the general language form. The overall presentation is that multiple clues are combined to determine the determination result, and the description of each clue conforms to the natural language sequence form. For example, in the auxiliary diagnosis and prediction project in the medical field, a variety of disease clues in the electronic medical record text are combined to determine the auxiliary diagnosis result, and the description of each condition is in the form of a natural language sequence. In this application, the clue that determines the result of the joint combination is the main information in the project-related text, and the related description information of the main information is the modification information of the main information.
[0080] Since the project-related text often contains multiple text units, such as multiple subsections, multiple paragraphs, etc., the information contained in different parts has different emphasis. Therefore, in this application, you can use specific text units in the project-related text Obtain the main information in the related text of the project and the decoration information of each main information. For example, in the electronic medical record, the "main complaint" and "current medical history" fields contain the main basis information for the doctor to make a diagnosis. The electronic medical record text can be obtained from the "main complaint" and "current medical history" in the electronic medical record text The main information in the main information, and the modification information of each main information.
[0081] For ease of understanding, in this application, assuming that the project-related text is an electronic medical record text, the main information in the electronic medical record text can be the examination category, examination result, organ and tissue name, symptom description, drugs, drugs, diseases, etc. The words of information such as surgery and examination methods, and the relevant description words of each word as the main information in the electronic medical record text are the modification information of the main information.
[0082] For example, the text of an electronic medical record contains the following information:
[0083] "History of present illness: Neck and shoulder pain started 2 years ago, accompanied by difficulty in raising arms, occasional dizziness, nausea, and headaches worsened. Shoulder and back pain, when the pain is severe, I feel irritable and cannot sleep. So I came to the hospital for treatment."
[0084] In the above electronic medical record text, shoulder and back pain, neck and shoulder pain, dizziness, nausea, headache, and irritability are the main information, severe is the modified information of shoulder and back pain, 2 years ago, difficult to raise the arm is the modified information of neck and shoulder pain, Modified information aggravated to headaches.
[0085] S103: Based on the main information and the modification information of each main information, construct a graph structure of the project-related text.
[0086] Because the graph structure has rich topological relationships, based on the nodes in the graph and the edges between nodes, it can represent a variety of clues in the project-related text, and it can also represent the description information of each clue in the project-related text. The clues become fuller and more specific, with finer granularity. Therefore, in this application, the graph structure of the project-related text can be constructed based on the backbone information and the modification information of each backbone information. The graph structure includes the nodes corresponding to the backbone information and the nodes corresponding to the modification information. There is an edge between the nodes.
[0087] In order to understand the graph structure more vividly, this application provides the graph structure of the electronic medical record text corresponding to the example in S102, specifically as figure 2 Shown.
[0088] S104: Analyze the graph structure of the item-related text, and determine the determination result of the item to be determined.
[0089] In this application, the diagram structure of the project-related text can indicate the main information of the project-related text and the modification information of each main information. By analyzing the diagram structure of the project-related text, it can be determined that the project-related text is used to determine The clues of the judgment result of the item to be judged and the description information of each clue, based on the clue and the description information of each clue, the judgment result of the item to be judged can be determined.
[0090] such as, figure 2 By analyzing the structure of the diagram, we can determine clues such as shoulder and back pain, neck and shoulder pain, dizziness, nausea, headache, irritability, etc. Severe is the description of shoulder and back pain. Two years ago, the difficulty of raising the arm was The description information of neck and shoulder pain and the description information of headache aggravated. Based on the above clues and the description information of each clue, the auxiliary diagnosis prediction result of the corresponding electronic medical record text can be given as "cervical spondylopathy".
[0091] This application discloses a project determination method. First, obtain the project-related text of the project to be determined, and then determine the main information in the project-related text and the modification information of each main information; based on the main information and each main information The modification information of the project determines the graph structure of the project-related text; finally, the graph structure of the project-related text is analyzed to determine the judgment result of the item to be judged. In this application, the method of analyzing project-related texts to achieve project determination can save manpower and time compared to manual determination, thereby reducing costs and improving project determination efficiency. Based on the above method, it is possible to realize the judgment of items such as auxiliary diagnosis prediction items in the medical field and auxiliary verification value prediction items in the judicial field.
[0092] Furthermore, using the project judgment method provided in this application can reduce subjective assumptions compared to manual judgment methods, reduce the arbitrariness of project judgments such as auxiliary diagnosis prediction projects in the medical field, and auxiliary verification value prediction projects in the judicial field. Ensure that the project judgment results are more objective and accurate, and enhance the transparency of the project judgment results.
[0093] In this application, the method of determining the main information in the project-related text and the modification information of each main information is also introduced as follows.
[0094] Among them, the way to determine the main information in the project related text can be:
[0095] Use the main element extraction tool to extract the main elements in the project-related text, and then use the word segmentation tool (such as jieba, etc.) to segment the main elements to obtain the main information.
[0096] Among them, the backbone element extraction tool has the ability to determine the backbone elements in the project-related text. For example, the existing neural network model can be trained based on a large number of project-related texts marked with the backbone elements to obtain the backbone element extraction tool.
[0097] The way to determine the modification information of each backbone information can be:
[0098] Taking each main information in the item-related text as the center, extracting preset number word information from the left and right sides of the main information respectively as the modification information of each main information, the preset number can be determined based on different items Need to set different values, for example, the neighborhood size can be set to 3.
[0099] In this application, the specific implementation method of constructing the graph structure of the project-related text based on the main information of the project-related text and the modification information of each main information is also disclosed, which may include:
[0100] S201: Based on the backbone information, construct a global dependency graph of the project-related text.
[0101] In this application, the backbone information can be used as the nodes of the global dependency graph, and an edge is constructed between every two nodes.
[0102] For ease of understanding, based on the main information corresponding to the electronic medical record text example in S102, a global dependency graph of the electronic medical record text can be constructed such as image 3 Shown.
[0103] S202: Based on the modification information of each backbone information, construct a local dependency graph of each backbone information.
[0104] In this application, for each backbone information, the modification information of the backbone information can be further used as a node, and the node corresponding to the backbone information is connected to the node corresponding to each modification information through an edge to construct a local dependency graph of the backbone information. It should be noted that if the backbone information has no modification information, the local dependency graph of the backbone information is not constructed. If the backbone information has multiple modification information, the nodes corresponding to the multiple modification information are also connected by edges.
[0105] For ease of understanding, the main information of neck and shoulder pain corresponding to the electronic medical record text example in S102, the partial dependency graph is as Figure 4 Shown.
[0106] S203: Construct a graph structure of the project-related text based on the global dependency graph of the project-related text and the local dependency graph of each main information of the project-related text.
[0107] In this application, the global dependency graph of the project-related text and the local dependency graph of each main information of the project-related text can be combined to obtain the graph structure of the project-related text. Specifically, the backbone information can be used as the nodes of the global dependency graph, and an edge is constructed between every two nodes. For the backbone information with decoration information, the decoration information of the backbone information can be further used as a node, the node corresponding to the backbone information is connected to the node corresponding to each decoration information through an edge, and any two nodes corresponding to the decoration information are connected, Get the graph structure of the training text, and each edge has a weight.
[0108] Among them, the calculation method of the weight of the edge can be as follows:
[0109] Based on the co-occurrence analysis of the two nodes corresponding to the edge, the weight of the edge between the two nodes is calculated, and the weight of the edge between the two nodes is the weight of the corresponding edge.
[0110] Specifically, the first number of sliding windows included in the corpus is determined, the second number of sliding windows included in the corpus of the first node of the two nodes is included in the sliding window included in the corpus, and two sliding windows included in the corpus are included The third number of sliding windows of the second node in the node, and the fourth number of sliding windows containing two nodes in the sliding windows included in the corpus; according to the first number, second number, third number, and fourth number Quantity, calculate the weight of the edge between two nodes. The corpus may be a corpus corresponding to the item to be determined, and specifically may be a training text set of the item.
[0111] As an implementable manner, the weight of the edge between two nodes i and j can be calculated based on the following formula:
[0112]
[0113]
[0114]
[0115] Among them, #W represents the first number of sliding windows included in the corpus, #W(i) represents the second number of sliding windows that contain node i in the sliding windows included in the corpus, #W(j) represents the number of sliding windows included in the corpus The sliding window contains the third number of sliding windows of node j, and #W(i,j) represents the fourth number of sliding windows containing two nodes i, j in the sliding windows contained in the corpus. The result of PMI (Point-wise Mutual Information) is the weight of the edge between two nodes i and j.
[0116] In this application, a specific implementation method for analyzing the graph structure of the item-related text to determine the determination result of the item-related text is also disclosed, which may include:
[0117] The graph structure of the item-related text is input into the item judgment model, and the item judgment model outputs the judgment result of the item-related text. The item judgment model uses the graph structure of the training text as the training sample and is labeled with the training text set The item judgment result of is obtained from sample label training.
[0118] It should be noted that GCN (Graph Convolution Network, graph convolutional neural network) has the ability to capture multi-granularity features in the graph structure. Therefore, in this application, the training text of the item to be determined can be used to train the GCN to obtain the item judgment model.
[0119] In order to facilitate understanding, this application gives a detailed introduction to the training method of the project judgment model, which is specifically as follows:
[0120] S301: Obtain the training text of the item to be determined.
[0121] In this application, the training text of the item to be determined can be the original training text, or the training text obtained after preprocessing and filtering the original training text.
[0122] As an implementable manner, the training text of the item to be determined in this application may specifically be a text format that meets the requirements of the item determination method, can successfully extract the main information, and annotated the regularized item determination result training text.
[0123] In order to understand the regularity of project judgment results, the following examples are given in this application:
[0124] "Lower left pneumonia", "lower right pneumonia", "left pneumonia" and "right pneumonia" are collectively referred to as "pneumonia";
[0125] "Acute bronchitis", "acute bronchiolitis", "acute bronchiolitis", "bronchiolitis", "chronic bronchitis", etc. are collectively referred to as "bronchitis".
[0126] S302: Construct a graph structure of the training text.
[0127] In this application, for each training text, the graph structure of the training text can be constructed in the following manner: obtain the main information in the training text, the modification information of each main information; based on the main information in the training text, each A graph structure of the training text is constructed with the modification information of the main information, and the graph structure of the training text includes the node corresponding to the main information and the node corresponding to the modification information.
[0128] When constructing the graph structure of the training text, you can first construct a global dependency graph of the training text based on the backbone information in the training text, and then construct the backbone information with modified information in the training text to construct each with modified information Finally, based on the global dependency graph of the training text and each local dependency graph, the graph structure of the training text is obtained.
[0129] Specifically, the backbone information can be used as the nodes of the global dependency graph, and an edge is constructed between every two nodes. For the backbone information with decoration information, the decoration information of the backbone information can be further used as a node, the node corresponding to the backbone information is connected to the node corresponding to each decoration information through an edge, and any two nodes corresponding to the decoration information are connected, Get the graph structure of the training text, and each edge has a weight. Among them, the calculation method of the weight of the edge can refer to the calculation method of the weight of the side of the project-related text, which will not be repeated here.
[0130] S303: Input the graph structure of the training text into the GCN, train the parameters of the GCN, and obtain the item judgment model based on the trained GCN.
[0131] It should be noted that, according to the nodes and edges in the graph structure of the training text, a node feature matrix and adjacency matrix are constructed, and the node feature matrix and adjacency matrix are used to input the GCN parameters to train the GCN parameters to obtain the item judgment model.
[0132] For ease of understanding, suppose the graph structure of the training text is G=(V,E), V={V 1 ,V 2 ,V 3 ,V 4 ,V 5 ,V 6 ,V 7 },E={(V 1 ,V 2 ) 10 ,(V 1 ,V 3 ) 2 ,(V 3 ,V 4 ) 2 ,(V 3 ,V 6 ) 11 ,(V 2 ,V 5 ) 1 ,(V 4 ,V 5 ) 4 ,(V 4 ,V 6 ) 6 ,(V 5 ,V 7 ) 7 ,(V 6 ,V 7 ) 3 } (Note: the data in the lower right corner of the vertex pair represents the weight of the edge), V represents the set of nodes in the graph structure G, E represents the set of edges in the graph G structure, the set of edges contains the corresponding to each edge The weight of nodes and edges.
[0133] Construct the node feature matrix of the training text according to the nodes in the graph structure G. Assuming that the number of nodes in the graph structure G is n, the node feature matrix is ​​X ∈ R nxm , Where m is the dimension of each node feature, and each row x v ∈R m Represents the feature vector of node v.
[0134] Construct the adjacency matrix A of the training text according to the set of edges in the graph structure G. The values ​​in the adjacency matrix are used to indicate the weights of the edges in the graph G, based on the formula Preprocess the adjacency matrix to get the preprocessed adjacency matrix Among them, D represents the degree value matrix, in the degree value matrix, there are only values ​​on the diagonal, and the value on the diagonal of each row represents the degree value of the node corresponding to the row, that is, the edge of the corresponding node of the row and other nodes Quantity.
[0135] Assuming that the GCN includes two graph convolutional layers, the weight information W of the two graph convolutional layers is randomly initialized 0 And W 1 , And preprocess the adjacency matrix of each training text And feature matrix X as the input data of the first layer of graph convolutional layer, the output of the first layer of graph convolutional layer is The output of the second layer graph convolutional layer is Among them, ρ is the activation function, such as ReLU, softmax, etc. After GCN converges, the optimal weight information W of the two-layer graph convolutional layer can be obtained 0 And W 1 , The weight information of the two-layer graph convolutional layer W 0 And W 1 The GCN that is the optimal weight information is the trained GCN.
[0136] Based on the above, this application also discloses a specific implementation method for determining and outputting the determination result of the item-related text by the item determination model, which is specifically as follows:
[0137] S401: The item determination model constructs a node feature matrix and an adjacency matrix according to the nodes and edges in the graph structure of the item-related text.
[0138] For the specific implementation of this step, refer to the method for constructing the node feature matrix and the adjacency matrix of the training text, which will not be repeated in this step.
[0139] S402: The item judgment model uses the node feature matrix and the adjacency matrix to obtain and output the judgment result of the item related text.
[0140] In this application, the project judgment model can obtain multiple judgment results and the probability of each judgment result, and finally output the judgment result with the highest probability as the judgment result of the project-related text.
[0141] For example, based on figure 2 The graph structure shown gives two results of "cervical spondylopathy" and "anxiety disorder". The probability of "cervical spondylosis" is 80%, and the probability of "anxiety disorder" is 60%, and the final judgment result is "Cervical Spondylopathy".
[0142] As an implementable manner, using the node feature matrix and the adjacency matrix to obtain and output the judgment result of the item-related text includes: using the node feature matrix and the adjacency matrix to obtain the features of each node; The feature of the node is merged to obtain the feature of the item related text, and based on the feature of the item related text, the judgment result of the item related text is obtained and output.
[0143] However, since each node plays a different role in item determination, as another possible implementation manner, using the node feature matrix and the adjacency matrix to obtain and output the determination result of the item-related text includes: After the node feature matrix and the adjacency matrix are obtained, the weight of each node is determined, and the feature of each node is weighted based on the weight to obtain the new feature of each node; Feature fusion is used to obtain the feature of the item-related text, and based on the feature of the item-related text, the judgment result of the item-related text is obtained and output.
[0144] Among them, the weight of each node can be calculated in the following way:
[0145] Determine the confidence of the node in the corpus, and the TF-IDF (Term Frequency—Inverse Document Frequency) of the node in the corpus; based on the confidence of the node in the corpus, and the node in the corpus In TF-IDF, determine the weight of the node.
[0146] As an implementable manner, the weight of the node can be calculated based on the following formula:
[0147] weight t, d =belief t, D *tfidf(t, d, D)
[0148] Among them, believe t, D Is the confidence of the node in the corpus, t represents the node, d is the project-related text, D represents the corpus, and tfidf(t, d, D) is the TF-IDF of the node in the corpus.
[0149] Specifically, the method for determining the confidence of the node in the corpus may be: determining the number of occurrences of the node in the corpus, and the number of occurrences of all nodes in the project-related text in the corpus, according to the occurrence of the node in the corpus The number of times, and the number of times that all nodes in the project-related text appear in the corpus, determine the confidence of the node in the corpus.
[0150] As an implementable manner, the confidence of the node in the corpus can be calculated based on the following formula:
[0151]
[0152] Among them, believe t,D Is the confidence level of the node in the corpus, t represents the node, D represents the corpus, t’ represents any node in the project-related text, f t, D Indicates the number of occurrences of the node in the corpus, ∑ t′ f t′, D Indicates the number of occurrences of all nodes in the project-related text in the corpus.
[0153] Specifically, the method for determining the TF-IDF of the node in the corpus can be: determining the number of occurrences of the node in the project-related text, the number of occurrences of all nodes in the project-related text, the number of corpora in the corpus, and the corpus containing the The number of the corpus of a node is determined according to the number of occurrences of the node in the project-related text, the number of occurrences of all nodes in the project-related text, the number of corpora in the corpus, and the number of corpora containing the node in the corpus to determine that the node is in the corpus TF-IDF in.
[0154] As an implementable manner, the TF-IDF of the node in the corpus can be determined based on the following formula.
[0155]
[0156] Where f t, d Indicates the number of occurrences of the node in the associated text of the project, ∑ t′∈d f t′, d Represents the number of occurrences of all nodes in the related text of the item, N represents the number of corpus in the corpus, n t Indicates the number of corpus that contains the node in the corpus.
[0157] Based on the above, this application provides a schematic diagram of obtaining an auxiliary diagnosis prediction result based on an auxiliary diagnosis prediction model, such as Figure 5 As shown, the electronic medical record text reads "Main complaint: recent severe shoulder and back pain, affecting normal life. Current medical history: neck and shoulder pain started 2 years ago, accompanied by difficulty raising arms, occasional dizziness, nausea, and headaches. Shoulder and back pain, severe pain I feel irritable and can’t sleep. So I came to the hospital for treatment.” The electronic medical record text was constructed as a graph structure, and the graph structure was analyzed through the auxiliary diagnosis prediction model, and the auxiliary diagnosis prediction result was cervical spondylosis.
[0158] The item determination device disclosed in the embodiment of the present application is described below, and the item determination device described below and the item determination method described above can be referred to each other.
[0159] Reference Image 6 , Image 6 This is a schematic diagram of the structure of an item determination device disclosed in an embodiment of this application. Such as Image 6 As shown, the item determination device may include:
[0160] The item-related text obtaining unit 11 is used to obtain the item-related text of the item to be determined;
[0161] The text information determining unit 12 is used to determine the main information in the project-related text and the modification information of each main information;
[0162] The project-related text diagram structure construction unit 13 is configured to construct the diagram structure of the project-related text based on the main information and the modification information of each main information;
[0163] The analysis unit 14 is configured to analyze the graph structure of the item-related text to determine the determination result of the item to be determined.
[0164] Optionally, the text information determining unit includes a modification information acquiring unit;
[0165] The modification information determining unit is configured to take each main information in the item-related text as the center, and extract preset number word information from the left and right sides of the main information respectively as the modification information of each main information.
[0166] Optionally, the project-associated text graph structure construction unit includes:
[0167] A global dependency graph construction unit for constructing a global dependency graph of the project-related text based on the main information;
[0168] A local dependency graph construction unit for constructing a local dependency graph of each backbone information based on the modification information of each backbone information;
[0169] The construction unit is configured to construct the graph structure of the project-related text based on the global dependency graph of the project-related text and the local dependency graph of each main information of the project-related text.
[0170] Optionally, the analysis unit includes:
[0171] The item determination model processing unit is used to input the graph structure of the item-related text into the item determination model, the item determination model outputs the determination result of the item-related text, and the item determination model is based on the graph structure of the training text The training sample is obtained by training the sample label with the judgment result of the item marked in the training text set.
[0172] Optionally, the device further includes a training text graph structure construction unit, the training text graph structure construction unit is configured to obtain the main information in the training text and the modification information of each main information; based on the training text The main information in, and the modification information of each main information, construct the graph structure of the training text.
[0173] Optionally, the item determination model processing unit includes:
[0174] A construction unit for constructing a node feature matrix and an adjacency matrix according to the nodes and edges in the graph structure of the project-related text;
[0175] The determination result determination unit is used to obtain and output the determination result of the item related text by using the node feature matrix and the adjacency matrix.
[0176] Optionally, the determination result determining unit is specifically configured to:
[0177] Using the node feature matrix and the adjacency matrix to obtain the feature of each node;
[0178] Obtaining the weight of each node in the associated text of the item;
[0179] Weighting the feature of each node based on the weight to obtain the new feature of each node;
[0180] Fuse the new features of the respective nodes to obtain the features of the associated text of the item;
[0181] Based on the characteristics of the item-related text, a determination result of the item-related text is obtained and output.
[0182] Optionally, the item to be determined is an auxiliary prediction item, the item-related text information of the item to be determined includes the related text of the auxiliary prediction item, and the item determination result of the item to be determined is the prediction result.
[0183] It should be noted that the specific function implementation of the above-mentioned units has been described in detail in the method embodiment, and will not be repeated in this embodiment.
[0184] Image 6 The hardware structure block diagram of the project determination device disclosed in the embodiments of this application, refer to Image 6 The hardware structure of the item determination device may include: at least one processor 1, at least one communication interface 2, at least one memory 3 and at least one communication bus 4;
[0185] In the embodiment of the present application, the number of the processor 1, the communication interface 2, the memory 3, and the communication bus 4 is at least one, and the processor 1, the communication interface 2, and the memory 3 communicate with each other through the communication bus 4;
[0186] The processor 1 may be a central processing unit CPU, or a specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present invention;
[0187] The memory 3 may include a high-speed RAM memory, or may also include a non-volatile memory (non-volatile memory), for example, at least one disk memory;
[0188] Wherein, the memory stores a program, and the processor can call the program stored in the memory, and the program is used for:
[0189] Get the project-related text of the project to be determined;
[0190] Determine the main information in the related text of the project and the modification information of each main information;
[0191] Based on the main information and the modification information of each main information, constructing the graph structure of the project-related text;
[0192] Analyze the graph structure of the item-related text to determine the determination result of the item to be determined.
[0193] Optionally, the detailed functions and extended functions of the program may refer to the above description.
[0194] The embodiments of the present application also provide a storage medium, which can store a program suitable for execution by a processor, and the program is used for:
[0195] Get the project-related text of the project to be determined;
[0196] Determine the main information in the related text of the project and the modification information of each main information;
[0197] Based on the main information and the modification information of each main information, constructing the graph structure of the project-related text;
[0198] Analyze the graph structure of the item-related text to determine the determination result of the item to be determined.
[0199] Optionally, the detailed functions and extended functions of the program may refer to the above description.
[0200] Finally, it should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply these entities Or there is any such actual relationship or sequence between operations. Moreover, the terms "include", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes those that are not explicitly listed Other elements of, or also include elements inherent to this process, method, article or equipment. If there are no more restrictions, the element defined by the sentence "including a..." does not exclude the existence of other same elements in the process, method, article, or equipment that includes the element.
[0201] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments can be referred to each other.
[0202] The above description of the disclosed embodiments enables those skilled in the art to implement or use this application. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined in this document can be implemented in other embodiments without departing from the spirit or scope of the application. Therefore, this application will not be limited to the embodiments shown in this text, but should conform to the widest scope consistent with the principles and novel features disclosed in this text.
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

no PUM

Description & Claims & Application Information

We can also present the details of the Description, Claims and Application information to help users get a comprehensive understanding of the technical details of the patent, such as background art, summary of invention, brief description of drawings, description of embodiments, and other original content. On the other hand, users can also determine the specific scope of protection of the technology through the list of claims; as well as understand the changes in the life cycle of the technology with the presentation of the patent timeline. Login to view more.
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Similar technology patents

10kv live-line work insulating support rod

InactiveCN103606844ASave manpower and timeEnsure job safetyApparatus for overhead lines/cablesScrew jointLive-line working
Owner:STATE GRID CORP OF CHINA +1

Gene variation pathogenic grade determination method and device

InactiveCN106951730ASave manpower and timeAccurate divisionBiostatisticsProteomicsMonogenic inheritanceDecision-making
Owner:为朔医学数据科技(北京)有限公司

Classification and recommendation of technical efficacy words

  • Save manpower and time
  • Low cost

Leather cutting machine

ActiveCN101586171AAccelerateSave manpower and timeLeather clicking/perforating/clickingComputer engineeringPunching
Owner:DONGGUAN EMMA CNC TECH CO LTD

System and method for transmitting wireless digital service signals via power transmission lines

ActiveUS7929940B1Reduce bandwidth requirementLow costInterconnection arrangementsWireless systems/telephoneElectric power transmissionBase transceiver station
Owner:NEXTEL COMMUNICATIONS

Nanowire structures comprising carbon

ActiveUS20070212538A1Low costImproved electrical connectivity and stabilityElectric discharge heatingFuel cells groupingNanometreMembrane configuration
Owner:ONED MATERIAL INC
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products