A semantic analysis method for crowdsourced design resources

Through semantic analysis methods, crowdsourcing design resources are divided into independent functional components and multi-level semantic models are constructed, which solves the problem of differential management of design resources on crowdsourcing platforms, realizes structured processing and unified modeling of resources, and improves search and Matching efficiency.

CN114970543BActive Publication Date: 2024-11-01OCEAN UNIV OF CHINA
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210543747.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-05-19
Publication Date
2024-11-01
Estimated Expiration
2042-05-19

AI Technical Summary

Technical Problem

Due to the self-organization characteristics of crowdsourcing design resources on the Internet crowdsourcing platform, there are great differences in the relevant content between design resources and it is difficult to manage, and they need to be structured.

Method used

A semantic analysis method for crowdsourcing service resources is proposed. Through phrase division and dependency analysis, crowdsourcing design resources are divided into independent functional components, and a multi-level semantic model is constructed to convert unstructured natural language short sentence descriptions into Structured set of relationships.

Benefits of technology

It realizes unified modeling of crowdsourcing design resources, which is of great significance to subsequent retrieval and matching, and improves the efficiency of resource management.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114970543B_ABST
    Figure CN114970543B_ABST
Patent Text Reader

Abstract

The invention discloses a semantic analysis method for crowdsourcing design resources, comprising: step 1, word separation and part-of-speech tagging of crowdsourcing design resource short sentences; step 2, segmenting the short sentences processed in step 1 into a plurality of independent short sentences; step 3, performing the following processing on each independent short sentence: dependency analysis, extracting independent functional components in the independent short sentence by parallel relationship and independent structure, and constructing a multi-level semantic model for each independent functional component; the invention segmentes crowdsourcing design resource short sentences, performs part-of-speech tagging, divides independent short sentences, divides independent functional components for each independent short sentence, and constructs a multi-level semantic model for each independent functional component, thereby converting unstructured natural language short sentence descriptions into structured relationship sets, realizing unified modeling of crowdsourcing design resources, and having important significance for subsequent retrieval and matching.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of computer data processing, and in particular, relates to a semantic analysis method for crowdsourcing design resources. Background Art

[0002] Crowdsourced design resources on Internet crowdsourcing platforms are mostly expressed in a mixture of text, numbers, and image data. The self-organizing characteristics of crowdsourcing services lead to great differences in the relevant content among design resources. To manage design resources well, it is necessary to structure the relevant service content involved. Summary of the invention

[0003] The present invention proposes a semantic analysis method for crowdsourcing service resources. The design resources on the Internet crowdsourcing platform are taken as the research object. Through phrase division and dependency analysis, the crowdsourcing design resources are divided into independent functional components, and multi-level semantic representation is performed on the independent functional components, thereby converting unstructured natural language short sentence descriptions into structured relationship sets to achieve unified modeling of crowdsourcing design resources.

[0004] The present invention is implemented by the following technical solutions:

[0005] A semantic analysis method for crowdsourced design resources is proposed, including:

[0006] Step 1: Separate words and tag parts of speech for the crowdsourcing design resource sentences;

[0007] Step 2, dividing the short sentence processed in step 1 into multiple independent short sentences;

[0008] Step 3: Perform the following processing on each independent sentence:

[0009] 1) Dependency analysis;

[0010] 2) Extract independent functional components from independent sentences using parallel relationships and independent structures;

[0011] 3) Construct a multi-level semantic model for each independent functional component:

[0012] (1) Traverse the independent functional components to find the core verb-predicate, and obtain the primary semantics with the detected generalized verb-object relationship; the generalized verb-object relationship includes direct object relationship, indirect object relationship, prepositioned object relationship, the core verb-predicate and its dominated noun phrase as adverbial-predicate relationship, the core verb-predicate and its dominated noun phrase as subject-predicate relationship, and the modifier expressed as attributive-predicate relationship when the object of the core verb-predicate is missing;

[0013] (2) Detect the modified components of the predicate center word and object center word in the primary semantics to obtain the secondary semantics;

[0014] (3) Detect the modifying components of the central word of the secondary semantics to obtain the tertiary semantics;

[0015] (4) Complete the detected semantic components above level three onto the central word to which they belong.

[0016] Furthermore, in step 2, the short sentence is divided into multiple independent short sentences by using spaces, semicolons, commas, and forward and backward slashes.

[0017] Furthermore, independent functional components in independent sentences are extracted based on parallel relationships and independent structural relationships, including:

[0018] (1) When the detected parallel relationship is located in the generalized object modification area, it is equally divided into two independent functional components based on the core verb and predicate;

[0019] (2) When the detected parallel relationship is between the core verb and predicate, it is equally divided into two independent functional components;

[0020] (3) When a parallel relationship and an independent structure exist at the same time, if the independent structure is a verbal predicate and there is a parallel relationship, it is equally divided into two independent functional components based on the verbal predicate of the independent structure; if the independent structure is not a verbal predicate, no independent functional component is established.

[0021] Furthermore, a multi-level semantic model is constructed for each independent functional component, including:

[0022] When it is detected that the first-level semantics does not have a core object, its modified part is regarded as its object and built into the first-level semantics.

[0023] Compared with the prior art, the advantages and positive effects of the present invention are: in the semantic analysis method of crowdsourcing design resources proposed in the present invention, the crowdsourcing design resource sentences are segmented, part-of-speech tags are performed, independent sentences are divided, each independent sentence is divided into independent functional components, and a multi-level semantic model is constructed for each independent functional component, thereby converting the unstructured natural language sentence description into a structured relationship set, realizing the unified modeling of crowdsourcing design resources, which is of great significance for subsequent retrieval and matching.

[0024] Other features and advantages of the present invention will become more apparent after reading the detailed description of the embodiments of the present invention in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0026] Figure 1 A schematic diagram of the analysis steps of the semantic analysis method of crowdsourcing design resources proposed in the present invention;

[0027] Figure 2 This is one of the schematic diagrams for dividing independent short sentences in the present invention;

[0028] Figure 3 This is the second schematic diagram of dividing independent short sentences in the present invention;

[0029] Figure 4 The third diagram of the division of independent sentences based on the dependency analysis tree in the present invention;

[0030] Figure 5 This is one of the schematic diagrams for dividing independent functional components based on the dependency analysis tree in the present invention;

[0031] Figure 6 The second schematic diagram of the division of independent functional components based on the dependency analysis tree in the present invention;

[0032] Figure 7 The third diagram of the division of independent functional components based on the dependency analysis tree in the present invention;

[0033] Figure 8 The following is a schematic diagram of the steps for constructing a multi-level semantic model in the present invention;

[0034] Fig. 9 This is one of the schematic diagrams for constructing a multi-level semantic model in the present invention;

[0035] Fig.10 The second schematic diagram of constructing a multi-level semantic model in the present invention;

[0036] Fig.11 The third schematic diagram of constructing a multi-level semantic model in the present invention;

[0037] Fig.12 This is the fourth schematic diagram of constructing a multi-level semantic model in the present invention. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] In the description of the present invention, it should be understood that the terms "center", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship are based on the orientation or position relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0040] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected" and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0041] The terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "plurality" means two or more.

[0042] The semantic analysis method of crowdsourcing design resources proposed in this invention is as follows: Figure 1 As shown, including:

[0043] Step S1: Separate words and tag parts of speech of crowdsourced design resource sentences.

[0044] In Chinese, characters are the basic units of words. To understand the meaning of a phrase or a sentence, it is necessary to divide it into words. This requires that when a computer processes a text, it must first divide the sentence into words, that is, automatically identify each word and add boundary markers to separate the words.

[0045] In actual projects, the commonly used method is to use rule-based algorithms for word segmentation, and then use statistical methods to assist. This can efficiently and accurately segment text words, and can also take into account the recognition of new words and unrecorded vocabulary. This is the hybrid word segmentation method.

[0046] For example, the word segmentation result of the short sentence "various brush fonts writing, design" describing a crowdsourcing resource function is: [various, brush, font, writing,,, design].

[0047] After word segmentation, the parts of speech need to be marked. Part of speech tagging is to determine the part of speech of each word in the sentence and mark it, dividing each word into nouns, verbs, adjectives, auxiliary words, etc.

[0048] In Chinese, many parts of speech are not single, and different parts of speech are expressed in different sentences. On the other hand, most words tend to have only one or two parts of speech, and one of the two (or more) parts of speech is used more frequently than the others. There is no unified annotation standard for Chinese part-of-speech tagging. The two mainstream annotations are the Peking University part-of-speech tagging set and the Pennsylvania part-of-speech tagging set. The following are examples of part-of-speech tagging:

[0049] The part-of-speech classification result of “various brush fonts writing and designing” is: [various / rz, brush / n, font / n, writing / v, 、 / w, design / vn].

[0050] The word class markers involved are mainly divided into the following categories: 1) noun (n); 2) verb (v); 3) punctuation (w); 4) adjective (a); 5) conjunction (c); 6) pronoun (r); 7) quantifier (q); 8) other (o). More detailed divisions include: gerund (vn); intransitive verb (vi); food (nf); demonstrative pronoun (rz), etc.

[0051] Step S2: Divide the short sentence processed in step S1 into multiple independent short sentences.

[0052] Punctuation marks are very common in the design resource information on crowdsourcing websites, such as "I can design logo drawings, product appearance, etc.", "Red wine, food, real-life photography", "Homepage design / professional PS cutouts / watermark removal / advertising design / product description", and the functions of these punctuation marks are mostly to divide short sentences. Therefore, the present invention divides short sentences into multiple independent short sentences through spaces, semicolons, commas, forward and backward slashes, etc., and short sentences without punctuation marks do not need to be divided.

[0053] For example:

[0054] 1) Phrases separated by slashes: "Website construction / graphic design / Baidu ranking / Weibo marketing speaks with quality!", the segmentation results refer to Figure 2 shown.

[0055] 2) Phrases separated by spaces: "food dishes desserts drinks beverages soda photography still life product photography photography", the segmentation results refer to Figure 3 shown.

[0056] 3) Phrases separated by commas: "Undertake various brochure and menu layout design", the segmentation results are as follows Figure 4 shown.

[0057] Step S3: Perform the following processing on each independent short sentence:

[0058] 1. Dependency analysis.

[0059] A dependency relationship is a relationship between words, that is, a binary asymmetric relationship between a central word and its subordinate words. The central word of a sentence is usually a verb, and all other words depend on the central word.

[0060] The following syntactic dependency relations are mainly used in crowdsourcing resources: 1) Coordination relation (COO); 2) Compound noun (FM); 3) Other noun (QM); 4) Direct object (VOB); 5) Indirect object (IOB); 6) Fronted object (FOB); 7) Adverbial-verb relation (ADV); 8) Subject-verb relation (SBV); 9) Attributive-predicate relation (ATT); (10) Independent structure (IC).

[0061] The definitions of the various relationships are as follows:

[0062] The parallel relationship (COO) indicates the word relationship connected by "and, with, or", and is used for the next step of processing of its parallel nouns or verbs, such as "details page and homepage design": {(details page / n– >homepage / n, COO).

[0063] Compound nouns (FM) indicate a noun-modifying relationship and are used as objects to describe the objects of a verb after being split, such as "dishes, desserts and sweets": {(dishes / n– > desserts / n, FM); (desserts / n– > desserts / n, FM).

[0064] Other nouns (QM) indicate the relationship between other parts of speech other than nouns that modify nouns, describe the attributes of nouns, and are used for subclassification, such as "Mini program development": {(mini / a– > program / n, QM).

[0065] A direct object (VOB) refers to the direct object following the predicate verb. Both direct objects and preposed objects are classified as verb-object structures, which are the basic dependency relations in crowdsourcing corpora, such as "send her a bouquet of flowers": {(send / v -> flowers / n, VOB)}.

[0066] The preposed object (FOB) refers to the object before the predicate verb, such as "mini program development": {(program / n–development / v, FOB).

[0067] Indirect object (IOB) refers to a personal noun after the predicate verb. It occasionally appears in the crowdsourcing corpus, such as “send her a bouquet of flowers”: {(send / v -> she / n, IOB)}, but this structure is not very meaningful and will not appear in the final dependency relationship.

[0068] The subject-predicate relationship (SBV) represents the relationship between the doer and the action in a sentence. For example, "I can help you...": {(I / r -> help / v, SBV)}.

[0069] The adverbial-middle relationship (ADV) represents the relationship where a verb modifies another verb. For example, "I can help you...": {(can / v -> help / v, ADV)}.

[0070] The attributive-middle relationship (ATT) represents the relationship between an attribute and the head. For example, "system security": {(system / n -> security / adj, ATT)}.

[0071] The independent structure (IC) indicates that two independent semantic components are connected. For example, "website development website construction": {(website development -> website construction, IC)}.

[0072] 2. Extract the independent functional components in the independent short sentences based on the coordination relationship and the independent structure.

[0073] The functional description of crowdsourcing design resources can be short or long. A complex description usually contains multiple core business functions, which can also be called having multiple independent functional components. There is no punctuation separation between each component. For example: "penetration testing vulnerability scanning website security detection intrusion detection security testing", "mobile mall website construction online shopping online payment shopping website development", etc. This kind of description is obviously a collection of multiple business functions. For this kind of description, the present invention extracts the independent functional components of the independent short sentences through dependency analysis.

[0074] Specifically, based on dependency analysis, the independent functional components are divided according to the coordination relationship (COO) and the independent structure (IC). Since the positions of relationships such as COO and IC are different, different semantic relationships will be caused. The present invention divides the independent functional components according to the following three situations.

[0075] (1) When the detected coordination relationship (COO) is located in the general object modification area, it is equally divided into two independent functional components based on the core verb-predicate.

[0076] Taking "WeChat and system security testing" as an example, its dependency analysis tree is shown in Figure 5 , according to the semantic understanding, based on the core verb-predicate "testing", the coordinated objects "WeChat" and "system" and the core verb-predicate "testing" are equally divided into two independent functional components:

[0077] Independent functional component 1: (The construction of multi-level semantics will be described in detail in the following content)

[0078] First-level semantics: {(testing, security, (ATT))}

[0079] Secondary semantics: {(WeChat, security, (ATT))}

[0080] Independent functional ingredient 2:

[0081] First level semantics: {(test, security, (ATT))}

[0082] Secondary semantics: {(system, security, (ATT))}

[0083] (2) When the detected parallel relationship is between the core verb and predicate, it is equally divided into two independent functional components.

[0084] Take "penetration test, vulnerability scanning, website security detection, intrusion detection, security testing" as an example. This description is lengthy and unclear, but the dependency analysis tree can express the semantic logical relationship among them. Figure 6 The dependency analysis tree shown can detect three core verbs "test 1", "detection", and "test 2", where test 1 and 2 are the same name:

[0085] The two core verbs and predicates form a parallel relationship and are equally divided into two independent functional components. In this embodiment, three independent functional components are divided: (the construction content of multi-level semantics will be described in detail below)

[0086] Independent functional ingredient 1:

[0087] First-level semantics: {(test1, penetration, (ATT)), (test1, scan, (ATT))}

[0088] Secondary semantics: (vulnerability, scanning, (ATT))

[0089] Independent functional ingredient 2:

[0090] First-level semantics: {(detection, website, (ATT), (detection, intrusion, (ATT)), (detection, security, (ATT))}

[0091] Independent functional ingredients 3:

[0092] First level semantics: {(test2, security, (ATT))}

[0093] (3) When a parallel relationship and an independent structure exist at the same time, if the independent structure is a verbal predicate and there is a parallel relationship, it is equally divided into two independent functional components based on the verbal predicate of the independent structure; if the independent structure is not a verbal predicate, no independent functional component is established.

[0094] Taking "Shanghai company social security account opening professional registration bookkeeping and tax filing agent with rich experience and super value" as an example, based on Figure 7The dependency analysis tree shown can detect a core verb predicate "open an account" and four independent structures "register", "bookkeeping", "agent", "experienced" and "special offer". Among them, the independent structures "register", "bookkeeping" and "agent" are all verb predicates and are in parallel relationship. Then, "register", "bookkeeping" and "agent" are equally divided into three independent structures based on "register", "bookkeeping" and "agent", and a total of 4 independent functional components are divided. Among them, "experienced" and "special offer" are non-verbal predicates and no independent functional components are established:

[0095] Independent functional ingredient 1:

[0096] First-level semantics: {(opening an account, social security, (SBV))}

[0097] Secondary semantics: {(enterprise, social security, (ATT))}

[0098] Level 3 semantics: {(Shanghai Company, Enterprise, (ATT))}

[0099] Independent functional ingredient 2:

[0100] First-level semantics: {(register, ∅, (∅))}

[0101] Secondary semantics: {(professional, registered, (ADV))}

[0102] Independent functional ingredients 3:

[0103] First-level semantics: {(bookkeeping, ∅, (∅))}

[0104] Independent functional ingredients 4:

[0105] First-level semantics: {(agent, ∅, (∅))}

[0106] Secondary semantics: {(tax return, agency, (ADV))}

[0107] 3. Construct a multi-level semantic model for each independent functional component.

[0108] The present invention constructs a multi-level semantic model based on the following predefined:

[0109] The first-level semantics takes the verb predicate as the core word and the noun phrase as the subordinate word. It expresses the basic semantics of the sentence through direct object relationship (VOB), indirect object relationship (IOB), fronted object relationship (FOB), etc., and generally represents a main function provided by the service resources.

[0110] Based on the primary semantics, the secondary semantics modifies and supplements the respective central words through the adverbial-verb relationship (ADV), the verb-complement relationship (CMP), and the preposition-object relationship (POB).

[0111] Based on the secondary semantics, the tertiary semantics continues to modify and supplement the central words of the secondary semantics, making the semantics richer and fuller.

[0112] The natural language of service resources expressed in phrases can basically express the entire meaning of the resources through the third-level semantics. Grammatical components higher than the third level can be processed by truncation or completion to the third-level central word.

[0113] When constructing the first-level semantics, considering the errors in dependency analysis, many relationships that should be subject-object relationships from the actual semantics will be identified as attributive-mediate-text (ATT) relationships. Therefore, for the first-level semantics, if it is detected that the core verb predicate and its dominant noun phrase are in an ATT relationship, it is also regarded as a subject-object relationship. The core verb predicate and its dominant noun phrase are in a subject-verb (SBV) relationship, which is also regarded as a subject-object relationship. When the object of the core verb predicate is missing, it may be due to the expression of the language itself or the error of the dependency analysis model itself. In this case, the semantics is mainly expressed by the core predicate and its modified parts. In this application, the modification expressed in the attributive-mediate-text (ATT) relationship is also regarded as the object and is located at the first-level position.

[0114] Therefore, when constructing the primary semantics, the present invention application refers to the direct object relationship, indirect object relationship, prepositioned object relationship, core verb-predicate relationship with its dominated noun phrase as adverbial-predicate relationship, core verb-predicate relationship with its dominated noun phrase as subject-predicate relationship, and modifiers expressed as attributive-predicate relationship when the object of the core verb-predicate is missing as generalized verb-object relationship. The reason for this design is that in both cases, the dominated noun can be regarded as the patient role, and can be combined with the verbal predicate to fully express the semantics, thereby obtaining a semantic expression model that is as general as possible.

[0115] Based on the above predefined method, the present invention proposes a method for constructing a multi-level semantic model, such as Figure 8 As shown, including:

[0116] 1. Traverse the independent functional components to find the core verb-predicate, and obtain the primary semantics based on the generalized verb-object relationship detected.

[0117] The broad subject-verb relationship here includes direct object relationship, indirect object relationship, preposed object relationship, adverbial-predicate relationship between the core subject-verb and the noun phrase it controls, subject-predicate relationship between the core subject-verb and the noun phrase it controls, and modifiers expressed in attributive-predicate relationship when the object of the core subject-verb is missing.

[0118] 2. Detect the modifying components of the predicate center word and object center word in the primary semantics to obtain the secondary semantics.

[0119] The predicate core word and object core word are modified and supplemented through the adverbial-verb relationship (ADV), verb-complement relationship (CMP), and preposition-object relationship (POB).

[0120] 3. Detect the modifying components of the central word of the secondary semantics to obtain the tertiary semantics.

[0121] The central word of the secondary semantics is different from the predicate central word and the object central word in the primary semantics. It is the modifying and supplementing part of the two. If there are no other modifying words in the modifying and supplementing part, the tertiary semantics will not be constructed.

[0122] 4. Complete the semantic components higher than the third level onto the central word to which they belong.

[0123] The entire extraction process is performed by traversing the dependency analysis tree from bottom to top in the opposite direction of the relationship arcs of the dependency analysis tree.

[0124] Taking "self-made handmade skin care products, top-grade makeup remover oil" as an example, its dependency analysis tree is shown in Fig. 9 , the core verb predicate "self-made" and its object "makeup remover" constitute the first-level semantics. In the second-level semantics, the predicate center word "self-made" has no modified part. The modified parts of the object center word "makeup remover" "top" and "skin care products" form the second-level semantics. The center word "skin care products" in the second-level semantics has a modified part, so the third-level semantics is constructed with "skin care products" as the center word. The multi-level semantic model obtained according to the above steps is as follows:

[0125] First-level semantics: {(homemade, makeup remover oil, (VOB))}

[0126] Secondary semantics: {(top level, cleansing oil, (ATT)), (skin care products, cleansing oil, (ATT))}

[0127] Level 3 semantics: {(handmade, skin care products, (ATT))}

[0128] It can be seen that the first-level semantics reflects the main function of the service resource, while the second- and third-level semantics are the modification parts of the core word.

[0129] In actual applications, although the resource description of a single independent functional component has only one core business function, it can be divided into the following two situations:

[0130] (1) The existence of a core object at the first level

[0131] Taking "renting temporary network space at a low price" as an example, the syntax tree obtained through dependency analysis is as follows Fig.10 As shown, its multi-level semantic model is given as:

[0132] First-level semantics: {(rental, network space, (VOB))}

[0133] Secondary semantics: {(low price, rental, (ADV)), (temporary, network space, (ATT))}

[0134] (2) There is no core object at the first level

[0135] As mentioned above, the missing object of the core verb predicate may be due to the expression of the language itself or the error of the dependency analysis itself. In this case, the semantics is mainly expressed by the core predicate and its modified parts. In this invention, the modification expressed by the attributive-central (ATT) relationship is still regarded as the object and is located in the position of primary semantics.

[0136] Taking "website security detection" as an example, its dependency analysis tree is as follows Fig.11 As shown, its multi-level semantic model is:

[0137] First level semantics: {(detection, security, (ATT))}

[0138] Secondary semantics: {(website, security, (ATT))}

[0139] Based on the semantic analysis method of crowdsourcing design resources proposed in the present invention, the crowdsourcing design resource sentences are segmented, part-of-speech tagged, divided into independent sentences, each independent sentence is divided into independent functional components, and a multi-level semantic model is constructed for each independent functional component, thereby converting the unstructured natural language sentence description into a structured relationship set, realizing the unified modeling of crowdsourcing design resources, which is of great significance for subsequent retrieval and matching.

[0140] Finally, taking "mobile phone mall website construction, online shopping, online payment and shopping website development" as an example, the semantic analysis method of crowdsourcing design resources proposed in this invention is applied to obtain a structured relationship set, and the dependency relationship analysis tree is shown in FIG. Fig.12 , structured processing includes:

[0141] 1. Segment and tag the sentence.

[0142] 2. Split the short sentence processed in step 1 into multiple independent short sentences using spaces, semicolons, commas, and forward and backward slashes;

[0143] This short sentence has no spaces, semicolons, commas, or forward and backward slashes, so this step can be skipped and the whole sentence can be considered as an independent short sentence.

[0144] 3. Perform dependency analysis on the independent sentence.

[0145] 4. Extract the independent functional components in the independent sentence by using parallel relationship and independent structure.

[0146] By traversing the independent sentence, we retrieved that its core verbs and predicates are "construction" and "development". According to the parallel relationship (COO), we divide it into two independent functional components with "construction" and "development" as the core verbs and predicates: independent functional component 1 (the core verb and predicate is "construction") and independent functional component 2 (the core verb and predicate is "development").

[0147] There are no independent structures such as independent verbs and independent nouns in this independent sentence, so there is no need to divide the independent functional components according to the independent structures.

[0148] 5. Construct a multi-level semantic model for each independent functional component.

[0149] For independent functional component 1, we get its first-level semantics {(construction, website 1, (ATT))} based on the generalized subject-object relationship. We search for the modified parts of the predicate center word "construction" and the object center word "website" in the first-level semantics to get the second-level semantics {(mall, website 1, (ATT))}. We search for the modified parts of the center word "mall" in the second-level semantics to get the third-level semantics {(mobile phone, mall, (ATT))}. The multi-level semantic model is as follows:

[0150] Independent functional ingredient 1:

[0151] First-level semantics: {(construction, website 1, (ATT))}

[0152] Secondary semantics: {(mall, website 1, (ATT))}

[0153] Level 3 semantics: {(mobile phone, mall, (ATT))}

[0154] For independent functional component 2, we get its first-level semantics {{(development, website 2, (ATT))} based on the generalized subject-object relationship. We search for the modified parts of the predicate center word "development" and the object center word "website" in the first-level semantics to get the second-level semantics {{(online shopping, website 2, (ATT)), (shopping, website 2, (ATT))}. We retrieve the modified parts of the center word "shopping" in the second-level semantics to get the third-level semantics {{(online payment, shopping, (ADV))}. The multi-level semantic model is as follows:

[0155] Independent functional ingredient 2:

[0156] First-level semantics: {(development, website2, (ATT))}

[0157] Secondary semantics: {(online shopping, website 2, (ATT)), (shopping, website 2, (ATT))}

[0158] Level 3 semantics: {(online payment, shopping, (ADV))}

[0159] The above-mentioned websites 1 and 2 have the same name but are distinguished from each other.

[0160] It should be noted that the dependency analysis tree mentioned in the above content of the present invention is another existing analysis means for semantic analysis, which is not a technical means limited by the present invention. The semantic analysis method of the present invention can be supported by adopting existing methods.

[0161] It should be noted that, in the specific implementation process, the above-mentioned control part can be implemented by a hardware processor executing computer execution instructions in software form stored in the memory, which will not be elaborated here, and the programs corresponding to the actions performed by the above-mentioned control circuit can be stored in the computer-readable storage medium of the system in software form, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

[0162] The computer-readable storage medium mentioned above may include volatile memory, such as random access memory; may also include non-volatile memory, such as read-only memory, flash memory, hard disk or solid-state drive; may also include a combination of the above types of memory.

[0163] The processor mentioned above may also be a general term for multiple processing elements. For example, the processor may be a central processing unit, or other general-purpose processors, digital signal processors, application-specific integrated circuits, field programmable gate arrays or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc., or may be a dedicated processor.

[0164] It should be pointed out that the above description is not a limitation of the present invention, and the present invention is not limited to the above examples. Changes, modifications, additions or substitutions made by ordinary technicians in this technical field within the essential scope of the present invention should also fall within the protection scope of the present invention.

Claims

1. A semantic analysis method for crowdsourced design resources, It is characterized in that include: Step 1: Separate words and tag parts of speech for the crowdsourcing design resource sentences; Step 2, dividing the short sentence processed in step 1 into multiple independent short sentences; Step 3: Perform the following processing on each independent sentence: 1) Dependency analysis; 2) Extract independent functional components from independent sentences using parallel relationships and independent structures, including: (1) When the detected parallel relationship is located in the generalized object modification area, it is equally divided into two independent functional components based on the core verb and predicate; (2) When the detected parallel relationship is between the core verb and predicate, it is equally divided into two independent functional components; (3) When a parallel relationship and an independent structure exist at the same time, if the independent structure is a verbal predicate and there is a parallel relationship, it is equally divided into two independent functional components based on the verbal predicate of the independent structure; if the independent structure is not a verbal predicate, no independent functional component is established. 3) Construct a multi-level semantic model for each independent functional component: (1) Traverse the independent functional components to find the core verb-predicate, and obtain the primary semantics with the detected generalized verb-object relationship; the generalized verb-object relationship includes direct object relationship, indirect object relationship, prepositioned object relationship, the core verb-predicate and its dominated noun phrase as adverbial-predicate relationship, the core verb-predicate and its dominated noun phrase as subject-predicate relationship, and the modifier expressed as attributive-predicate relationship when the object of the core verb-predicate is missing; (2) Detect the modified components of the predicate center word and object center word in the primary semantics to obtain the secondary semantics; (3) Detect the modifying components of the central word of the secondary semantics to obtain the tertiary semantics; (4) Complete the detected semantic components above level three onto the central word to which they belong.

2. The semantic analysis method of crowdsourced design resources according to claim 1, It is characterized in that In step 2, the short sentence is divided into multiple independent short sentences by using spaces, semicolons, commas, and forward and backward slashes.

3. The semantic analysis method of crowdsourcing design resources according to claim 1, It is characterized in that A multi-level semantic model is constructed for each independent functional component, including: When it is detected that the first-level semantics does not have a core object, its modified part is regarded as its object and built into the first-level semantics.