And further classifying the sentence entities belonging to the element texts by using a pipeline model. Classifying the sentence entities belonging to the element text into any one of industry scale, development history and trend, policy and regulation and competitive format, and specifically, referring to table 2. Other categories may also include in the classification process: the classification process of the categories, such as industry barriers, business models, industry drivers, core indicators, financing events, node definitions, and life cycles, may refer to the above classification process, and will not be described herein in detail. It is to be understood that the type of element text obtained by recognition may be one or more, and is not limited herein.

TABLE 2

In recognition of a sentence entity belonging to a sentence of the element text, a second task is performed using the pipeline model used in the above step 203 as the model. It can be understood that the classification labels preset in the pipeline model and the classification criteria and examples mentioned in the table are provided for explaining the method, and may be adjusted according to the requirements of the method in the actual implementation process of the method, without limiting the method.

205. And identifying the sentence entity by using a sequence labeling model to obtain a plurality of industry chain nodes and the relationship attributes among the industry chain nodes.

And identifying the sentence entity by using a sequence labeling model to obtain a plurality of industrial chain nodes. And identifying the obtained sentence entities by using a sequence marking model so as to further obtain the industrial chain nodes in the target text, wherein the sequence marking model can identify industrial chain elements including but not limited to industrial entities, weight entities and the like in the text, and if a new industry is found, outputting the industrial chain elements as a new node in the industrial chain. And further obtaining a plurality of industry chain nodes included in the target text. The relationship attribute between a plurality of industry chain nodes, that is, the influence weight relationship between different industry chain nodes, is noteworthy in that the relationship attribute between the industry chain nodes can be identified and obtained by the sequence marking model, but the specific corresponding relationship between the relationship attribute and the industry chain nodes cannot be identified and obtained by the sequence marking model, for example, after the target text is identified based on the sequence marking model, the relationship attribute D, the relationship attribute E, and the relationship attribute F between the industry chain nodes are obtained, and the relationship attribute D, the relationship attribute E, or the relationship attribute F between the industry chain node a and the industry chain node B can be obtained by further identifying the relationship attribute D, the relationship attribute E, or the relationship attribute F.

It can be understood that, before the sequence labeling model is used, the model should be trained using a corresponding text as a training set, so that the trained sequence labeling model has the capability of identifying an industrial chain node, which is not described herein in detail. The neural network Model used in The process of identifying The industry chain node may be any one of sequence labeling models, such as Hidden Markov Models (HMMs), Maximum Entropy models (The Maximum Entropy models), and Conditional Random Fields (CRFs), which are not specifically limited herein, and The neural network Model used may be structurally adjusted to improve The identification effect of The industry chain node, i.e., a variant of The sequence labeling Model is used to identify industry chain elements such as The industry chain node and The weight entity, which is not specifically limited herein.

206. And setting the industrial chain node key words corresponding to the industrial chain link points based on the industrial chain nodes.

And setting the industrial chain node key words corresponding to the industrial chain link points based on the industrial chain nodes. The specific manual setting by the staff can be performed by a neural network model with the function. Specifically, the worker may determine the related vocabulary as the industrial keyword according to the name of the industrial chain node, or the industrial keyword may be obtained through the frequency of occurrence of the vocabulary or the importance degree of the vocabulary in the text by other neural networks, which may be determined according to the actual situation, and is not limited herein.

207. And identifying the sentence entities by using a relation extraction model to obtain the upstream and downstream and inclusion relations among the plurality of industrial chain nodes.

207. And identifying the sentence entities by using a relation extraction model to obtain the upstream and downstream and inclusion relations among the plurality of industrial chain nodes. And automatically realizing upstream and downstream relation recognition based on the relation extraction model, further obtaining upstream and downstream relations and inclusion relations of a plurality of industrial chain link points, and automatically connecting different nodes by combining an industrial chain map in a directional arrow form.

208. And using a relation extraction model to correlate the industrial chain nodes and the relation attribute corresponding relation among the industrial chain nodes, and obtaining the relation attribute corresponding relation among the industrial chain nodes and the industrial chain nodes.

Using a relationship extraction model to correlate the industrial chain nodes and the relationship attribute corresponding relationship between the industrial chain nodes to obtain the relationship attribute corresponding relationship between the industrial chain nodes and the industrial chain nodes, and specifically using a weight dictionary to participate in the relationship attribute setting, such as

Setting the weight dictionary includes: most important, generally, less influential, etc. And realizing the corresponding relation between the influence weight in the target text and the industrial entity pair based on the deep learning model.

Example 1:

text: the injection molding machine is wide downstream and has small periodic fluctuation, so that the injection molding machine can keep steady growth in the future, and the market scale reaches 400 hundred million in 2015. At present, household appliances and automobiles are the largest downstream, and medical treatment and 3C are growth bright spots.

Upstream: injection molding machine

Downstream: household electrical appliance and automobile

Weight coefficient: maximum of

The weight type: description of the nature

The weight direction: downstream

Example 2:

text: at present, the silicon wafer accounts for about 33% of the whole photovoltaic cell cost, because financial crisis reduces subsidies in new energy industries of various countries, the yield is huge and difficult to digest, the profit of the silicon wafer industry is rapidly reduced after 2009, and more importance is attached to cost control.

Upstream: silicon wafer

Downstream: photovoltaic cell

Weight coefficient: 33 percent

The weight type: cost quantification

The weight direction: downstream

It is understood that, in the identification process, the obtained qualitative influence weight expression in the form of important, very important, etc. is also converted into a numerical description in the form of percentage, such as maximum conversion into 0.8, very important conversion into 0.6, important conversion into 0.4, etc., which may be specific to the actual situation and is not limited herein.

And associating the industrial chain nodes and the relationship attributes between the industrial chain nodes based on the relationship extraction model, wherein the corresponding relationship attribute between the industrial chain node A and the industrial chain node B is an influence weight F. It should be noted that the relationship extraction model used in the step 207 may be the same model but different algorithms or training methods, and is not limited herein.

209. Using a neural network model to identify the sentence entity, and obtaining the representative company and the financing situation

And identifying the sentence entities by using a neural network model based on deep learning to obtain the typical company and the financing condition, and adding the obtained typical company and the financing condition to an industrial chain node as part of attribute information of the industrial chain node so as to improve the expressive content of the industrial chain model. The concrete section adds the website of the homepage of the typical company and the corresponding description to the corresponding information of the industrial chain node.

210. And filling the industrial chain model based on the identification result to obtain an industrial chain map.

And attribute information of the industrial chain nodes obtained based on the steps is as follows: the key words of the industry chain nodes, the element texts, the typical company and financing situation, and the element texts comprise: industry scale, development history and trends, policy and regulation, competition patterns and the like are filled in the industry chain node part in the industry chain model, and the obtained relationships among the industry chain nodes are identified to comprise upstream and downstream relationships among the industry chain nodes and/or influence weight relationships among the industry chain nodes. Adding the obtained data to an industrial chain model to obtain a complete industrial chain map, wherein the specific obtained industrial chain map can be referred to fig. 3.

In the implementation process of the present embodiment, the process of obtaining the influence weight portion may perform a further detailed execution process to obtain more accurate influence weight data, specifically referring to fig. 4, including:

401. and establishing an influence weight dictionary among different industry chain nodes.

And establishing an influence weight dictionary among different industry chain nodes.

402. Determining relationship attributes between the plurality of industry chain nodes based on the impact weight dictionary.

The above steps 401 to 402 are similar to the implementation process of step 208 in the embodiment corresponding to fig. 2 of the present application, and are not described herein again.

404. And judging whether influence weight relations exist between all the industry chain link points with upstream and downstream or inclusion relations.

And judging whether influence weight relations exist among all the industrial chain link points with upstream and downstream or inclusion relations, if not, executing the step 405 to complete the influence weight relations among the industrial chain nodes with upstream and downstream or inclusion relations.

405. And completing the influence weight relationship between the industrial chain nodes with upstream and downstream or inclusion relationship.

And completing the influence weight relationship between the industrial chain nodes with upstream and downstream or inclusion relationship. That is, the method performs weight completion on the industry chain nodes which are identified to have upstream and downstream or inclusion relations but are not identified to obtain the influence weight attributes, and the specific completion mode includes qualitative mode and quantitative mode.

For the quantitative weight attribute, it specifically means that the attribute such as "cost ratio" or "demand ratio" of the two industries in the text is obtained. And acquiring the attributes and adding the corresponding values into the continuous edges to be used as continuous edge attributes. For the upstream and downstream industry chain nodes which cannot acquire the attributes, corresponding influence weight values can be set by practitioners based on experience, possible values can be obtained by calculation in a numerical calculation mode, known shares can be deducted, and the influence weight values can be obtained by averaging the remaining values.

For qualitative impact weight attributes. The three types of the most important, important and general can be manually set by a practitioner, and the attributes of the types can be added to the connecting edge after being digitalized and taken as the attributes of the connecting edge. The specific situation may be determined according to actual circumstances, and is not limited herein.

It can be understood that after the influence weight attribute is obtained, the above steps can be repeated according to real-time data to realize dynamic adjustment of the weight. In order to obtain real-time valid impact weight data. By the method, the effectiveness of the influence weight data in the industry chain map is improved.

After the industry chain map is obtained, the names of the industry chain nodes included in the industry chain map can be correspondingly processed, so that the names of the industry chain nodes are more accurate and standard, please refer to fig. 5 specifically.

501. And acquiring a standard industrial chain node system.

And acquiring a standard industrial chain node system, wherein the standard industrial chain node system comprises a plurality of standard industrial chain nodes. Firstly, a standard industry node system needs to be established, and industry chain nodes included in the standard industry node system are a set obtained after industry researchers research national economy industry and industry attributes. The method is characterized in that a Standard Industry node system is established based on existing data and Standard Classification systems such as national economy Industry Classification, national input and output Table and Global Industry Classification Standard (GICS), wherein each Standard Industry node corresponds to a plurality of Industry chain nodes.

502. Performing text representation on the standard industry chain nodes and the industry chain nodes obtained based on target text recognition based on a text representation model, and obtaining the similarity between the standard industry chain nodes and the industry chain nodes obtained based on target text recognition

And performing text representation on the standard industry chain nodes and the industry chain nodes obtained based on the target text recognition based on a text representation model, and obtaining the similarity between the standard industry chain nodes and the industry chain nodes obtained based on the target text recognition. And processing and calculating the standard industrial chain nodes and the industrial chain nodes obtained by identification by using a text representation model to obtain the similarity between the industrial chain nodes obtained by identification and the standard industrial chain nodes.

And identifying the standard industry chain nodes and the industry chain nodes obtained based on the target text recognition based on a text representation model, and obtaining the similarity between the standard industry chain nodes and the industry chain nodes obtained based on the target text recognition. Firstly, a standard industry chain node system is analyzed by using a text representation model including but not limited to a thermal coding technology, a bag-of-words model, a pre-training language model and the like to obtain a standard industry vector matrix, then, the name of an industry chain node obtained based on target text recognition is input into the text representation model to obtain a representation vector of the industry chain node, and finally, similarity measurement indexes including but not limited to Cosine similarity (Cosine), Spearman's rank correlation coefficient (Spearman's rank correlation) and the like are used for calculating the similarity between the standard industry chain node and the standard industry chain node.

503. And judging whether the similarity between the industrial chain node obtained based on the target text recognition and the standard industrial chain node is greater than a preset threshold value.

And judging whether the similarity between the industrial chain node obtained based on the target text recognition and the standard industrial chain node is greater than a preset threshold value. If the value is greater than or equal to the preset threshold value, executing step 504, setting the industrial chain node obtained based on the target text recognition as a common industrial chain node, and corresponding the common industrial chain node with the standard industrial chain node. If the value is smaller than the preset threshold value, executing step 505, setting the industry chain node obtained based on the target text recognition as a demonstrative industry chain node, and adding the demonstrative industry chain node to the standard industry chain system. The preset threshold may be set according to actual situations, and is not limited herein.

504. And setting the industrial chain nodes obtained based on the target text recognition as common industrial chain nodes, and corresponding the common industrial chain nodes with the standard industrial chain nodes.

If the similarity between the industrial chain link obtained based on the target text recognition and the standard industrial chain node is greater than or equal to a preset threshold value, the node is correspondingly embodied in a standard industrial chain node system, so that the industrial chain node obtained based on the target text recognition can be set as a common industrial chain node, and the common industrial chain link corresponds to the standard industrial chain node. Specifically, labels may be added to note the specific corresponding standard industrial chain nodes, which is not described herein in detail.

505. And setting the industrial chain node obtained based on the target text recognition as a demonstrative industrial chain node, and adding the demonstrative industrial chain node to the standard industrial chain system.

If the similarity between the industrial chain node obtained based on the target text recognition and the standard industrial chain node is smaller than a preset threshold value, it indicates that the industrial chain node does not exist in a standard industrial chain system, the industrial chain node obtained based on the target text recognition can be set as an illustrative industrial chain node, and the illustrative industrial chain node is added to the standard industrial chain system. Specifically, the demonstrative industrial chain nodes and the common industrial chain nodes can be displayed in a distinguishing manner, so that a user can visually determine and analyze the demonstrative industrial chain nodes. It can be understood that, since the industry chain node obtained by the machine learning manner may have an error, the demonstrative industry chain node may be manually checked, and the erroneous industry chain node or the actually non-existent industry chain node may be deleted, so as to ensure the accuracy of the result, which is not limited herein.

The method for constructing an industry chain map in the application embodiment is described above, and the apparatus for constructing an industry chain map in the embodiment of the present invention is described below. Referring to fig. 3, an embodiment of the present application includes:

the establishing unit 601 is used for establishing an industry chain model;

the identification unit 602 is configured to identify a target text by using machine learning based on the industry chain model, and obtain an identification result related to the industry chain model;

a filling unit 603, configured to correspondingly fill the identification result into the industry chain model, so as to obtain an industry chain map.

In this embodiment, the flow executed by each unit in the industry chain chart building apparatus is similar to the flow of the industry chain chart building method described in the embodiment corresponding to fig. 1, and is not repeated here.

A server may include one or more Central Processing Units (CPUs) and memory having one or more applications or data stored therein.

In this embodiment, the specific functional module division in the central processing unit may be similar to the functional module division manner of each unit described in fig. 1, and is not described herein again.

Wherein the memory may be volatile storage or persistent storage. The program stored in the memory may include one or more modules, each of which may include a sequence of instructions operating on a server. Still further, a central processor may be provided in communication with the memory for executing a series of instruction operations in the memory on the server.

The server may also include one or more power supplies, one or more wired or wireless network interfaces, one or more input-output interfaces, and/or one or more operating systems, such as Windows Server, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, etc.

The central processing unit may perform the operations performed by the method for constructing an industry chain map in the embodiment shown in fig. 1, and details are not described herein.

The present invention also provides a computer-readable storage medium for implementing the functions of the industry chain graph building method, on which a computer program is stored, and when the computer program is executed by a processor, the processor can be used for executing the method as described in fig. 1.

It will be appreciated that the integrated units, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a corresponding one of the computer readable storage media or integrated as a computer program product for performing the above-described method. Based on such understanding, all or part of the flow of the method according to the above embodiments may be implemented by a computer program, which may be stored in a computer-readable storage medium and used by a processor to implement the steps of the above embodiments of the method. Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, usb disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM), Random Access Memory (RAM), electrical carrier wave signals, telecommunications signals, software distribution medium, and the like. It should be noted that the computer readable medium may contain content that is subject to appropriate increase or decrease as required by legislation and patent practice in jurisdictions, for example, in some jurisdictions, computer readable media does not include electrical carrier signals and telecommunications signals as is required by legislation and patent practice.

The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.

It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.

In the several embodiments provided in the present application, it should be understood that the disclosed system, apparatus and method may be implemented in other manners. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.

The above-mentioned embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims

1. An industrial chain map construction method is characterized by comprising the following steps:

establishing an industrial chain model;

2. The method of constructing an industry chain map according to claim 1,

3. The method of claim 2, wherein the industrial chain graph is constructed,

4. The method for constructing an industry chain map spectrum according to claim 3, wherein the identifying a target text based on the industry chain model to obtain an identification result related to the industry chain model comprises:

identifying the industrial chain nodes and sentence entities corresponding to the relationship attributes among the industrial chain nodes by using a relationship extraction model to obtain influence weight relationships among the industrial nodes;

5. The method according to claim 4, wherein the relationship attribute correspondence between the industry chain nodes and the industry chain nodes is obtained by associating the relationship attribute correspondence between the industry chain nodes and the industry chain nodes by using a relationship extraction model, and then the method further comprises:

6. The method of industry chain atlas construction of claim 5, the method further comprising:

7. The method of industry chain atlas construction of claim 1, the method further comprising:

8. An industrial chain map building apparatus, comprising:

the establishing unit is used for establishing an industrial chain model;

and the filling unit is used for correspondingly filling the identification result into the industry chain model to obtain an industry chain map.

9. An industrial chain map building apparatus, comprising:

the memory is a transient memory or a persistent memory;

the central processor is configured to communicate with the memory, the instructions in the memory being executable on the central processor to perform the method of any of claims 1-7.

10. A computer-readable storage medium comprising instructions that, when executed on a computer, cause the computer to perform the method of any one of claims 1-7.