Industrial design quality evaluation method and system

Abstract

The invention discloses an industrial design quality evaluation method and system. The method comprises the steps that enterprise industrial design quality is divided into a first evaluation domain and a second evaluation domain; determining the scale of each evaluation index in the first evaluation domain according to the independent completion degree; determining the scale of each evaluation index in the second evaluation domain according to the stage matching degree; collecting process files of enterprise industrial design, and attaching classification labels to the process files according to the evaluation indexes of the first evaluation domain and the second evaluation domain; analyzing the matching degree between the process file under each classification label and each scale of the corresponding evaluation index, and determining the scale with the high matching degree as a calibration score of each evaluation index; and respectively counting the calibrated total score of the first evaluation domain and the second evaluation domain. According to the method, the industrial design quality of an enterprise can be quantitatively evaluated, so that the current situation that the evaluation is carried out only depending on the subjective impression of an evaluator is avoided; observation points of the evaluation process are defined, a multi-dimensional industrial design quality evaluation system is provided, and the industrial pertinence is high.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial design, and in particular to an industrial design quality assessment method and system. Background Art

[0002] Industrial design is a key tool to promote the development of product manufacturing, and its role in the expansion of product markets and the upgrading of enterprises is becoming increasingly prominent. At the same time, with the gradual deepening of industrial design services, the maturity of industrial design units themselves is gradually increasing. From the perspective of the entire process of product development, industrial design participates in and serves the entire process from product definition, design, manufacturing, transportation and service. The more links involved and the higher the degree of participation, the greater the influence of industrial design.

[0003] In the entire manufacturing product development and design process, industrial design is a relatively independent unit. With the improvement and perfection of the product development system, the role of the industrial design sector has become increasingly prominent. How to objectively observe its role and quantify its impact. Making it possible for the industry to conduct a comprehensive quantitative evaluation of industrial design is a reliable path for the industry to influence and intervene in it. In particular, the decomposition of the design and development links and the quantitative analysis of the impact increase the maturity of the design profession, thereby improving the level of the entire design service.

[0004] The establishment of modern design majors began at the Bauhaus School of Design in Germany, and the discussion of design standards began during the period of the German Manufacturing Union. In particular, the teaching experiments of the Ulm College in Germany and the ten standards of "good design" proposed by Dieter Rams have established a basic framework for design evaluation. Later, internationally influential design competitions such as "IF" and "RED DOT" provided specific standards for design evaluation. Their standards are recognized by the international design community. They are also accepted by the market and enterprises. However, the observation points and influence of the existing evaluation methods lack demonstration and cannot be quantified. The judges are subjective and arbitrary, and the control points of scoring are not consistent. Summary of the invention

[0005] In view of the above problems existing in the prior art, the object of the present invention is to provide a more objective, accurate and quantifiable industrial design quality evaluation system.

[0006] In order to achieve the above-mentioned invention object, the technical solution provided by the present invention includes:

[0007] An industrial design quality assessment method, comprising:

[0008] Divide the enterprise's industrial design quality into the first evaluation domain and the second evaluation domain;

[0009] The first evaluation domain reflects the maturity of the enterprise's industrial design quality, including seven evaluation indicators: product definition, appearance, ergonomics, interactive interface, technical implementation, development process and brand establishment; the second evaluation domain reflects the focus of the enterprise's industrial design quality, including three evaluation indicators: problem awareness value, originality and design completion;

[0010] The scale of each evaluation indicator in the first evaluation domain is determined by the independent completion degree; the scale of each evaluation indicator in the second evaluation domain is determined by the stage matching degree;

[0011] Collecting process files of the enterprise's industrial design, and adding classification labels to the process files according to the evaluation indicators of the first evaluation domain and the second evaluation domain;

[0012] Analyze the matching degree between the process files under each classification label and the corresponding evaluation index scales, and determine the scale with the highest matching degree as the calibration score of each evaluation index;

[0013] The total calibration scores of the first evaluation domain and the second evaluation domain are calculated respectively.

[0014] In some preferred embodiments, the first evaluation domain also includes three additional scoring evaluation indicators: professional design awards, patent status and professional article publication.

[0015] In some preferred embodiments, a weighting coefficient of 20% is given to the three evaluation indicators of product definition, appearance and ergonomics in the first evaluation domain, which have high enterprise participation and great influence.

[0016] In some preferred embodiments, the method of analyzing the matching degree between the process files under each classification label and the corresponding evaluation index scales includes:

[0017] A long short-term memory recursive neural network based on a word-level attention mechanism is constructed and trained, the process file is input into the long short-term memory recursive neural network based on a word-level attention mechanism, and the process file is classified by scale matching degree using an evidence classifier.

[0018] In some preferred embodiments, before the process file is input into the long short-term memory recursive neural network based on the word-level attention mechanism, a preprocessing step is also included: using information extraction technology to extract a summary of the process file, and inputting the extracted summary as a text data set into the long short-term memory recursive neural network based on the word-level attention mechanism.

[0019] The present invention also provides an industrial design quality evaluation system, comprising:

[0020] collecting a classification module, a first evaluation module, a second evaluation module and an output module;

[0021] The collection and classification module is configured to collect enterprise industrial design process files, add classification labels to the process files according to the evaluation indicators of the first evaluation domain and the second evaluation domain, and pass them to the first evaluation module and the second evaluation module;

[0022] The first evaluation module is configured to include seven first evaluation sub-units, namely, product definition, appearance, ergonomics, interactive interface, technical implementation, development process and brand establishment. Each of the first evaluation sub-units corresponds to a classification label and has a built-in scale determined according to the independent completion degree. The first evaluation sub-unit calculates the scale matching degree for the process files with the corresponding classification labels respectively, and determines the scale with the high matching degree as the calibration score of the corresponding evaluation index.

[0023] The second evaluation module is configured to include three second evaluation sub-units, namely, problem awareness value, originality and design completion, each of which corresponds to a classification label and has a built-in scale determined according to the stage matching degree; the second evaluation sub-unit calculates the scale matching degree for the process files with the corresponding classification labels respectively, and determines the scale with a high matching degree as the calibration score of the corresponding evaluation indicator;

[0024] The output module is configured to respectively summarize the calibrated total scores of the first evaluation module and the second evaluation module, and output the calibrated total score of the first evaluation module as an evaluation result reflecting the maturity of the enterprise's industrial design quality; and output the calibrated total score of the second evaluation module as an evaluation result reflecting the focus of the enterprise's industrial design quality.

[0025] In some preferred embodiments, the first evaluation module also includes three additional scoring evaluation sub-units: professional design awards, patent status and professional article publication.

[0026] In some preferred embodiments, the three first evaluation sub-units of product definition, appearance and ergonomics are weighted by adding 20% ​​to the calibration scores of the corresponding evaluation indicators.

[0027] In some preferred embodiments, the first evaluation subunit and the second evaluation subunit are both equipped with a built-in trained long short-term memory recursive neural network based on a word-level attention mechanism, and the long short-term memory recursive neural network based on a word-level attention mechanism uses an evidence classifier to classify the scale matching degree of process files corresponding to classification labels.

[0028] In some preferred embodiments, the first evaluation subunit and the second evaluation subunit are both equipped with a built-in preprocessing unit for extracting summaries of process files using information extraction technology, and inputting the extracted summaries as text data sets into a long short-term memory recursive neural network based on a word-level attention mechanism.

[0029] Beneficial Effects

[0030] 1. The present invention can evaluate the industrial design quality of an enterprise in a more objective, accurate and quantifiable manner, freeing the design evaluation from the current situation where it is solely based on the subjective impression of the evaluator; 2. It clarifies the scale of the evaluation process and proposes a multi-dimensional industrial design quality evaluation system, which is highly industry-specific and easy to operate. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 A flowchart of a method in a preferred embodiment of the present invention;

[0032] Figure 2 This is an example table of evaluation indicators of the first evaluation domain in a preferred embodiment of the present invention;

[0033] Figure 3 This is an example table of evaluation indicators of the second evaluation domain in a preferred embodiment of the present invention;

[0034] Figure 4 A system structure diagram of a preferred embodiment of the present invention; DETAILED DESCRIPTION

[0035] In order to make the purpose, technical solution and advantages of the present invention clearer, the present invention is further described below in conjunction with the accompanying drawings. In the description of the present invention, it should be understood that the terms "upper", "lower", "front", "back", "left", "right", "top", "bottom", "inside", "outside" and the like indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings, which are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention.

[0036] Example 1

[0037] like Figure 1 As shown, this embodiment provides an industrial design quality assessment method, comprising the steps of:

[0038] S1. Divide the enterprise's industrial design quality into the first evaluation domain and the second evaluation domain;

[0039] The first evaluation domain reflects the maturity of the enterprise's industrial design quality, including seven evaluation indicators: product definition, appearance, ergonomics, interactive interface, technical implementation, development process and brand establishment; among them, the seven evaluation indicators in the first evaluation domain represent the degree of participation of the enterprise's industrial design team in the product design process. An enterprise with a mature industrial design team must have a deeper degree of participation, including:

[0040] The product definition refers to the description and definition of its basic elements at the beginning of product development. It is one of the basic tasks of industrial design. Compared with traditional industrial design tasks, today industrial design has more and more say in product definition in the entire product development, and the judgment of the participation of industrial design in the product definition process is an important part of judging the influence of industrial design.

[0041] The appearance design refers to the system design involving the product appearance. It belongs to the basic task scope of traditional industrial design. In enterprises with excellent industrial design quality, it should be led by the industrial design team, rather than by other teams (such as independent appearance creativity or outsourcing teams).

[0042] The ergonomics refers to the study of the interaction and reasonable combination of people, machines and the environment. The ergonomics technology of excellent industrial design should be completed by the relevant professionals under the leadership of the industrial design team.

[0043] The interactive interface refers to the design and research of the relationship between the equipment and the operator in the interactive communication between people and equipment. Under the development trend of intelligent products, the proportion of interactive interface design is getting higher and higher.

[0044] The technical realization mentioned above refers to the optimization of technology, materials, etc. by industrial design. It is an important aspect of industrial design to expand design services, reflecting the design service capabilities and depth of the enterprise's industrial design team.

[0045] The development process refers to the scope of industrial design in the entire development process of major equipment manufacturing enterprises. In the past, the focus of industrial design was on the appearance design of product design. With the increasing requirements for product development, the scope of industrial design in the entire development process is also getting wider and wider. It reflects the higher requirements for industrial design in the development process of major equipment products.

[0046] Brand building refers to the systematic design of a certain type of product around the product brand. It is different from the limitation of traditional industrial design that only optimizes and transforms specific products. Instead, it builds the brand of the product from a systematic perspective, establishes the brand or enhances the brand value.

[0047] The above evaluation indicators reflect the maturity of the enterprise's industrial design quality. In some better embodiments, in order to refer to the third party's evaluation of industrial design products, according to their importance and objectivity, the first evaluation domain also includes three additional evaluation indicators: professional design awards, patent status and professional article publication.

[0048] Among them, the professional design awards refer to the design awards that focus on industrial design. The review standards and review processes of various industrial design awards have strict academic quality and good reputation. If a product wins an award, it is a good professional endorsement. The level of awards reflects the level of design.

[0049] The patent situation refers to whether the product design has been granted patents and the categories of the granted patents. It is an important indicator of the originality of the design, and its recognition of innovation is persuasive.

[0050] The professional article publication refers to the publication of theoretical or practical summary articles related to design in professional journals or media. Generally speaking, there is a positive correlation between the level of professional journals and the quality of design.

[0051] In other preferred embodiments, in order to ensure the relative accuracy of the evaluation results. For example, in the evaluation model product model, the degree of innovation of the invention patent is higher than that of the appearance patent, and there is also a difference in the weighted coefficients of the influence of the two. In some preferred embodiments, considering the influence of the industrial design focus, the top rankings are the appearance design of the product, the ergonomics of the product, and the participation and influence of the product definition part are the greatest, and the participation and influence of other parts are second. Therefore, a weighted coefficient of 20% is given to these three evaluation indicators. The weighted coefficient is set to distinguish the influence of different evaluation points in the same evaluation range by assigning a weighted coefficient to the influence of the evaluation point.

[0052] The second evaluation domain reflects the focus of the enterprise's industrial design quality, including three evaluation indicators: problem awareness value, originality and design completion;

[0053] Among them, the three evaluation indicators in the second evaluation domain represent the focus trend and influence of the enterprise industrial design team in the product design process. The higher the three evaluation indicators of an enterprise with a mature industrial design team, the better. Specifically, they include:

[0054] The problem awareness value refers to the value of the product determined by the problem awareness value in the product definition. However, problems are hierarchical, from high to low, they are problems faced by the entire human race, social group problems, personal problems, etc. The higher the level of problem solving, the higher the social value of the product.

[0055] The originality refers to the innovation of the product, which is also one of the important indicators to measure the value of the product and a key indicator to evaluate the design quality. The higher the originality, the higher the value of the design technology.

[0056] The degree of design completion refers to the review status of the degree of design, which corresponds to the three levels of multiple batch production and batch production prototype production. It is an important indicator for testing the maturity of design. The higher the level, the higher the degree of product completion.

[0057] It should be understood that other assessment points may also be included when evaluating the quality of industrial design, such as policy orientation correspondence, financial evaluation and other options. Since they are not the focus of protection of the present invention, they will not be elaborated here.

[0058] S2. Determine the scale of each evaluation indicator in the first evaluation domain based on the independent completion degree; determine the scale of each evaluation indicator in the second evaluation domain based on the stage matching degree; Specifically, those skilled in the art can divide each evaluation indicator into several scales (observation points) according to its characteristics and core demands, such as Figure 2 , Figure 3 As shown in the table example, observation points of different degrees correspond to different industrial design qualities. It should be understood that the setting of the scale is to quantify the assignment process of each evaluation indicator to avoid the influence of subjective factors on the evaluation process. Specifically, the grading method of the scale can be designed with reference to the average degree of participation of enterprises in this field in industrial design, or it can be set through statistical methods such as questionnaires, or it can be set with reference to selection standards and criteria with high recognition in the domestic and international industries. It should be understood that the focus of the scale of each evaluation indicator in the first evaluation domain is the independent completion of the enterprise in the process of completing the industrial design, and the focus of the scale of each evaluation indicator in the second evaluation domain is the stage matching of the enterprise in the process of completing the industrial design.

[0059] S3. Collect the process files of the enterprise's industrial design, and add classification labels to the process files according to the evaluation indicators of the first evaluation domain and the second evaluation domain; wherein the process files refer to the files generated by the enterprise in the entire life cycle of the industrial design, and have different formats and file contents according to the different stages of industrial design. For example, the enterprise needs to hold a meeting to discuss the direction and guidance of this industrial design at the beginning of the industrial design, so the process files of this stage may include meeting minutes, internal discussion drafts, etc.; the next step may involve the signing of a contract with a third-party cooperative developer, and the process files at this time may include contracts, agreements, etc.; with the specific development of industrial design, it will involve design drawings, production drawings, process manuals and other documents for specific product realization. It should be understood that the collection of process files and the addition of classification labels can be done manually or by converting physical files into electronic files and then by computer. The classification label is the classification corresponding to the evaluation indicators of the first evaluation domain and the second evaluation domain. It should be understood that the classification label of the same process file is not unique. For example, the classification labels of the development contract signed with a third-party software developer may include: "Human-machine engineering", "Interactive interface", and "Originality". Therefore, the addition of labels is to make the subsequent evaluation process of the process file more targeted, rather than an "either-or" binary multiple-choice question.

[0060] S4, analyzing the matching degree between the process files under each classification label and the corresponding evaluation index scales, and determining the scale with the highest matching degree as the calibration score of each evaluation index. Since most process files are text files formed by natural language, the scale matching can be performed by a person skilled in the art subjectively, or the process files can be processed by natural language and then the matching degree can be calculated by a classifier. For the latter, this embodiment provides a feasible implementation plan, which specifically includes:

[0061] A long short-term memory recursive neural network based on a word-level attention mechanism is constructed and trained, the process file is input into the long short-term memory recursive neural network based on a word-level attention mechanism, and the process file is classified by scale matching degree using an evidence classifier.

[0062] The word-level attention mechanism is based on the following considerations:

[0063] Natural language processing is divided into three levels according to the degree of attention to details during processing: word level (or granularity), word level (or granularity) and sentence level (or granularity). The higher the level, the higher the corresponding degree of attention to details, and the more details contained in the text can be extracted. However, this often also brings more occupation of computing resources. Considering the purpose of the present invention, the extraction of keywords of the process file can often obtain the substantive content reflected by the process file. For example, the development content agreed in the design contract signed with the third-party appearance designer can include keywords that clearly indicate the degree of third-party participation to prove the independent completion of the enterprise in the evaluation index of "appearance design". For example, keywords such as "full power" and "complete" reflect that the enterprise has limited influence in the evaluation index of "appearance design". When the keywords contain the limited content of a specific part or part, it can be explained that the enterprise has partially adopted other appearance design schemes in the industrial design process of the product. Therefore, the present invention does not need to be too demanding on details, and therefore does not need to carry out a word-level (or granularity) attention mechanism or a multi-level attention mechanism to speed up the efficiency of evaluation and save example resources.

[0064] The use of evidence classifier is based on the following considerations:

[0065] Existing classifiers are mainly divided into three categories: 1. Decision tree classifier; providing an attribute set, the decision tree classifies the data by making a series of decisions based on the attribute set. This process is similar to identifying a plant by its characteristics; its advantages are simple structure and mature technology, but when used in the environment of the present invention, its accuracy is poor because the output of its result is an either-or binary problem, and the present invention is not such a simple binary when scaling and matching the key content reflected according to semantics. 2. Selection tree classifier; use technology similar to the decision tree classifier to classify data. Unlike the decision tree, the selection tree contains special selection nodes, and the selection nodes have multiple branches; its advantage is that more branches and selection nodes are designed to enhance the accuracy of classification, but like the decision tree classifier, the accuracy of its output results is poor. 3. Evidence classifier; classifies data by checking the possibility of a specific result occurring based on a given attribute; this type of classifier outputs a probability, but when applied in the environment of the present invention, the same process file has different matching probabilities in two adjacent scales, and the result output by the classifier based on the matching probability is obviously more credible. In some preferred embodiments, the matching probability can be used as a scale weight to further improve the accuracy of scale matching.

[0066] In some preferred embodiments, considering that the amount of process file data generated by enterprises during industrial design is huge, in order to reduce the amount of data and speed up the evaluation efficiency, it is considered to set a preprocessing step before inputting the process file into the long short-term memory recursive neural network based on the word-level attention mechanism, including: using information extraction technology to extract a summary of the process file, and inputting the extracted summary as a text data set into the long short-term memory recursive neural network based on the word-level attention mechanism.

[0067] Example 2

[0068] This embodiment is developed on the basis of the above-mentioned embodiment 1. Figure 4 As shown, this embodiment provides an industrial design quality assessment system, including:

[0069] collecting a classification module, a first evaluation module, a second evaluation module and an output module;

[0070] The collection and classification module is configured to collect enterprise industrial design process files, attach classification labels to the process files according to the evaluation indicators of the first evaluation domain and the second evaluation domain, and pass them to the first evaluation module and the second evaluation module; wherein the enterprise industrial design process files include text, drawings, and video process materials generated during the industrial design process, as well as design plan task books, design review meeting minutes, patent authorization documents, article publication documents, award-winning materials, product production, sales materials, etc.

[0071] The data collection and classification module may collect data by crawling the database, acquiring data in batches from the Internet, or manually inputting data.

[0072] The classification performed by the collection and classification module mainly refers to the classification according to the evaluation index categories involved in the single process file and the attributes of the data itself, so as to facilitate the subsequent assignment operation. It should be understood that the same industrial design data can contain multiple classification labels. For example, production drawings can be assigned three classification labels: "appearance modeling", "ergonomics", and "design completion", and invention patent authorization certificates can be assigned three classification labels: "brand establishment", "patent status" and "originality".

[0073] The first evaluation module is configured to include seven first evaluation sub-units, namely, product definition, appearance, ergonomics, interactive interface, technical implementation, development process and brand establishment. Each of the first evaluation sub-units corresponds to a classification label and has a built-in scale determined according to the independent completion degree. The first evaluation sub-unit calculates the scale matching degree for the process files with the corresponding classification labels respectively, and determines the scale with the high matching degree as the calibration score of the corresponding evaluation index.

[0074] The first evaluation module is configured to include seven evaluation sub-units, namely, product definition, appearance, ergonomics, interactive interface, technical implementation, development process and brand establishment. The above evaluation sub-units are assigned values ​​according to the enterprise industrial design data transmitted after classification by the collection module; it should be understood that the specific method of the assignment may be to preset the assignment standard in the first evaluation module, and the first evaluation module analyzes the received enterprise industrial design data and matches it with the assignment standard, and assigns the value with a higher matching degree to the corresponding evaluation sub-unit. Specifically, the assignment process may be performed by manual analysis and assignment, or by using an assignment model trained by deep learning. The specific method is not the focus of the present invention and will not be described in detail here.

[0075] The second evaluation module is configured to include three second evaluation sub-units, namely, problem awareness value, originality and design completion, each of which corresponds to a classification label and has a built-in scale determined according to the stage matching degree; the second evaluation sub-unit calculates the scale matching degree for the process files with the corresponding classification labels respectively, and determines the scale with a high matching degree as the calibration score of the corresponding evaluation indicator;

[0076] The output module is configured to respectively summarize the calibrated total scores of the first evaluation module and the second evaluation module, and output the calibrated total score of the first evaluation module as an evaluation result reflecting the maturity of the enterprise's industrial design quality; and output the calibrated total score of the second evaluation module as an evaluation result reflecting the focus of the enterprise's industrial design quality.

[0077] In some preferred embodiments, the first evaluation module also includes three additional scoring evaluation sub-units: professional design awards, patent status and professional article publication.

[0078] In some preferred embodiments, the three first evaluation sub-units of product definition, appearance and ergonomics are weighted by adding 20% ​​to the calibration scores of the corresponding evaluation indicators.

[0079] In some preferred embodiments, the first evaluation subunit and the second evaluation subunit are both equipped with a built-in trained long short-term memory recursive neural network based on a word-level attention mechanism, and the long short-term memory recursive neural network based on a word-level attention mechanism uses an evidence classifier to classify the scale matching degree of process files corresponding to classification labels.

[0080] In some preferred embodiments, the first evaluation subunit and the second evaluation subunit are both equipped with a built-in preprocessing unit for extracting summaries of process files using information extraction technology, and inputting the extracted summaries as text data sets into a long short-term memory recursive neural network based on a word-level attention mechanism.

[0081] The industrial design quality assessment method and system provided by the present invention is mainly a basic assessment tool for the current status assessment of the design industry and the formulation of relevant industry policies. The corresponding indicators are quantitatively described, and the results can be used to compare different industries as a whole, and to analyze and compare the influence of various focus points within the industry.

[0082] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for evaluating industrial design quality, characterized in that: include: Divide the enterprise's industrial design quality into the first evaluation domain and the second evaluation domain; The first evaluation domain reflects the maturity of the enterprise's industrial design quality, including seven evaluation indicators: product definition, appearance, ergonomics, interactive interface, technical implementation, development process and brand establishment; the second evaluation domain reflects the focus of the enterprise's industrial design quality, including three evaluation indicators: problem awareness value, originality and design completion; The scale of each evaluation indicator in the first evaluation domain is determined by the independent completion degree; the scale of each evaluation indicator in the second evaluation domain is determined by the stage matching degree; Collecting process files of the enterprise's industrial design, and adding classification labels to the process files according to the evaluation indicators of the first evaluation domain and the second evaluation domain; Analyze the matching degree between the process files under each classification label and the corresponding evaluation index scales, and determine the scale with the highest matching degree as the calibration score of each evaluation index; The total calibration scores of the first evaluation domain and the second evaluation domain are calculated respectively.

2. The industrial design quality assessment method according to claim 1, characterized in that: The first evaluation domain also includes three additional scoring evaluation indicators: professional design awards, patent status and professional article publication.

3. The industrial design quality assessment method according to claim 1 or 2, characterized in that: In the first evaluation domain, a weighting coefficient of 20% is given to the three evaluation indicators of product definition, appearance design and ergonomics, which have high corporate participation and great influence.

4. The industrial design quality assessment method according to claim 1, characterized in that: The method for analyzing the matching degree between the process files under each classification label and the corresponding evaluation index scales includes: A long short-term memory recursive neural network based on a word-level attention mechanism is constructed and trained, the process file is input into the long short-term memory recursive neural network based on a word-level attention mechanism, and the process file is classified by scale matching degree using an evidence classifier.

5. The industrial design quality assessment method according to claim 4, characterized in that: Before the process file is input into the long short-term memory recursive neural network based on the word-level attention mechanism, a preprocessing step is also included: using information extraction technology to extract a summary of the process file, and inputting the extracted summary as a text data set into the long short-term memory recursive neural network based on the word-level attention mechanism.

6. An industrial design quality assessment system, characterized in that: include: collecting a classification module, a first evaluation module, a second evaluation module and an output module; The collection and classification module is configured to collect enterprise industrial design process files, add classification labels to the process files according to the evaluation indicators of the first evaluation domain and the second evaluation domain, and pass them to the first evaluation module and the second evaluation module; The first evaluation module is configured to include seven first evaluation sub-units, namely, product definition, appearance, ergonomics, interactive interface, technical implementation, development process and brand establishment. Each of the first evaluation sub-units corresponds to a classification label and has a built-in scale determined according to the independent completion degree. The first evaluation sub-unit calculates the scale matching degree for the process files with the corresponding classification labels respectively, and determines the scale with the high matching degree as the calibration score of the corresponding evaluation index. The second evaluation module is configured to include three second evaluation sub-units, namely, problem awareness value, originality and design completion, each of which corresponds to a classification label and has a built-in scale determined according to the stage matching degree; the second evaluation sub-unit calculates the scale matching degree for the process files with the corresponding classification labels respectively, and determines the scale with a high matching degree as the calibration score of the corresponding evaluation indicator; The output module is configured to respectively summarize the calibrated total scores of the first evaluation module and the second evaluation module, and output the calibrated total score of the first evaluation module as an evaluation result reflecting the maturity of the enterprise's industrial design quality; and output the calibrated total score of the second evaluation module as an evaluation result reflecting the focus of the enterprise's industrial design quality.

7. The industrial design quality assessment system according to claim 6, characterized in that: The first evaluation module also includes three additional scoring evaluation sub-units: professional design awards, patent status and professional article publication.

8. The industrial design quality assessment system according to claim 6 or 7, characterized in that: The three first evaluation sub-units of product definition, appearance and ergonomics are weighted by adding 20% ​​to the calibration scores of the corresponding evaluation indicators.

9. The industrial design quality evaluation system according to claim 6, characterized in that: The first evaluation subunit and the second evaluation subunit are both equipped with a trained long short-term memory recursive neural network based on a word-level attention mechanism. The long short-term memory recursive neural network based on a word-level attention mechanism uses an evidence classifier to classify the process files corresponding to the classification labels according to the scale matching degree.

10. The industrial design quality evaluation system according to claim 9, characterized in that: The first evaluation subunit and the second evaluation subunit are both equipped with a preprocessing unit for extracting summaries of process files using information extraction technology, and inputting the extracted summaries as text data sets into a long short-term memory recursive neural network based on a word-level attention mechanism.

Images