Engineering Consulting Service Evaluation Method and System

Abstract

Disclosed is a method and system for evaluating engineering consulting services. It first obtains multiple parameters of multiple processes of the entire process of the project to be evaluated, wherein the multiple parameters of each process include the actual completion period, the expected period, the actual investment and the expected investment. Next, a text description of the problems to be rectified of the project to be evaluated is obtained. Then, a feature interaction analysis is performed on the multiple parameters of the multiple processes and the text description of the problems to be rectified of the project to be evaluated to obtain the semantic interaction features of the project status and the problems to be rectified. Finally, based on the semantic interaction features of the project status and the problems to be rectified, the estimated duration for the problems to be rectified is determined. In this way, the subjectivity and inefficiency problems caused by manual evaluation can be avoided, and the mutual influence and correlation between each process link and each factor are utilized to provide a more accurate estimation of the rectification time consumption, providing strong support for engineering management.

Description

Technical Field

[0001] The present disclosure relates to the field of intelligent evaluation, and more specifically, to an engineering consulting service evaluation method and system thereof. Background Art

[0002] The purpose of engineering consulting service evaluation is to conduct a comprehensive evaluation and analysis of engineering projects in order to provide targeted consulting advice and decision-making support. It involves evaluating all aspects of engineering projects, including project planning, schedule management, cost control, quality assurance, and risk management. In the evaluation process of engineering consulting services, it is crucial to conduct a time-consuming evaluation of issues that need to be rectified. Evaluating the time consumption of rectification can help project managers develop reasonable project plans and schedules. By understanding the time and resource investment required for rectification, work processes and resource allocation can be reasonably arranged to ensure that rectification work is completed on time and avoid project delays and additional costs.

[0003] However, traditional evaluation schemes usually rely on the subjective judgment and experience of experts. Since the rectification process involves multiple process links and factors, different experts may give different evaluation results, resulting in high subjectivity and uncertainty in the evaluation results. In addition, there are complex correlations between the various links and factors in the rectification process. Traditional schemes that rely on manual experience judgment often find it difficult to fully consider the mutual influence and dependency between the various links, resulting in inaccurate evaluation results of time-consuming rectification.

[0004] Therefore, an optimized evaluation scheme for engineering consulting services is desired. Summary of the invention

[0005] In view of this, the present disclosure proposes an engineering consulting service evaluation method and system, which can avoid the subjectivity and inefficiency problems brought about by manual evaluation, and also utilize the mutual influence and correlation between various process links and various factors to provide a more accurate estimation of the time consumption for rectification, thereby providing strong support for engineering management.

[0006] According to one aspect of the present disclosure, there is provided a method for evaluating engineering consulting services, comprising: obtaining multiple parameters of multiple processes of the entire process of a project to be evaluated, wherein the multiple parameters of each process include an actual completion period, an estimated period, an actual investment, and an estimated investment; obtaining a text description of problems requiring rectification of the project to be evaluated; performing feature interaction analysis on the multiple parameters of the multiple processes and the text description of the problems requiring rectification of the project to be evaluated to obtain semantic interaction features between project status and problems requiring rectification; and determining an estimated duration for the problems requiring rectification based on the semantic interaction features between project status and problems requiring rectification.

[0007] According to another aspect of the present disclosure, there is provided an engineering consulting service evaluation system, comprising: a parameter acquisition module, for acquiring multiple parameters of multiple processes of the entire process of the project to be evaluated, wherein the multiple parameters of each process include the actual completion period, the expected period, the actual investment and the expected investment; a problem acquisition module for acquiring a text description of the problem to be rectified of the project to be evaluated; a feature interaction analysis module, for performing feature interaction analysis on the multiple parameters of the multiple processes and the text description of the problem to be rectified of the project to be evaluated to obtain the interaction features between the project status and the problem to be rectified semantics; and a duration estimation module, for determining the estimated duration for the problem to be rectified based on the interaction features between the project status and the problem to be rectified semantics.

[0008] According to the embodiment of the present disclosure, it first obtains multiple parameters of multiple processes of the entire process of the project to be evaluated, wherein the multiple parameters of each process include the actual completion period, the expected period, the actual investment and the expected investment, then obtains the text description of the problems to be rectified of the project to be evaluated, and then performs feature interaction analysis on the multiple parameters of the multiple processes and the text description of the problems to be rectified of the project to be evaluated to obtain the semantic interaction features between the project status and the problems to be rectified, and finally, based on the semantic interaction features between the project status and the problems to be rectified, determines the estimated duration for the problems to be rectified. In this way, the subjectivity and inefficiency caused by manual evaluation can be avoided, and the mutual influence and correlation between each process link and each factor are utilized to provide a more accurate estimation of the rectification time consumption, providing strong support for project management.

[0009] Further features and aspects of the present disclosure will become apparent from the following detailed description of exemplary embodiments with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate exemplary embodiments, features, and aspects of the disclosure and, together with the description, serve to explain the principles of the disclosure.

[0011] Figure 1 A flowchart of an engineering consulting service evaluation method according to an embodiment of the present disclosure is shown.

[0012] Figure 2 A schematic diagram showing the architecture of an engineering consulting service evaluation method according to an embodiment of the present disclosure.

[0013] Figure 3 A flowchart showing sub-step S130 of the engineering consulting service evaluation method according to an embodiment of the present disclosure is shown.

[0014] Figure 4A flowchart showing sub-step S131 of the engineering consulting service evaluation method according to an embodiment of the present disclosure is shown.

[0015] Figure 5 A flowchart showing sub-step S132 of the engineering consulting service evaluation method according to an embodiment of the present disclosure is shown.

[0016] Figure 6 A flowchart showing a training step further included in the engineering consulting service evaluation method according to an embodiment of the present disclosure.

[0017] Figure 7 A block diagram of an engineering consulting service evaluation system according to an embodiment of the present disclosure is shown.

[0018] Figure 8 An application scenario diagram of an engineering consulting service evaluation method according to an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present disclosure will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present disclosure, rather than all embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work also fall within the scope of protection of the present disclosure.

[0020] As shown in the present disclosure and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0021] Various exemplary embodiments, features and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0022] In addition, in order to better illustrate the present disclosure, numerous specific details are given in the following specific embodiments. It should be understood by those skilled in the art that the present disclosure can also be implemented without certain specific details. In some examples, methods, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the subject matter of the present disclosure.

[0023] In response to the above technical problems, the technical concept of the present disclosure is to collect data on multiple parameters of multiple processes in the entire engineering process, such as the actual completion period and expected period of the process, the actual input and expected input of the process, and obtain the problems that need to be rectified. Then, data processing and semantic understanding algorithms are introduced at the back end to perform correlation analysis of multiple parameters of multiple processes in the entire engineering process, as well as semantic understanding of the problems that need to be rectified, so as to automatically evaluate the time consumption of rectifying the problems. In this way, the subjectivity and inefficiency caused by manual evaluation can be avoided, and the mutual influence and correlation between each process link and each factor can be used to provide a more accurate estimation of the rectification time consumption, providing strong support for engineering management.

[0024] Figure 1 A flowchart of an engineering consulting service evaluation method according to an embodiment of the present disclosure is shown. Figure 2 FIG. 1 is a schematic diagram showing an architecture of an engineering consulting service evaluation method according to an embodiment of the present disclosure. Figure 1 and Figure 2 As shown, the engineering consulting service evaluation method according to the embodiment of the present disclosure includes the steps of: S110, obtaining multiple parameters of multiple processes of the entire process of the project to be evaluated, wherein the multiple parameters of each process include the actual completion period, the expected period, the actual investment and the expected investment; S120, obtaining a text description of the problems to be rectified of the project to be evaluated; S130, performing feature interaction analysis on the multiple parameters of the multiple processes and the text description of the problems to be rectified of the project to be evaluated to obtain the interaction features between the semantics of the project status and the problems to be rectified; and, S140, determining the estimated duration for the problems to be rectified based on the interaction features between the semantics of the project status and the problems to be rectified.

[0025] Specifically, in the technical solution disclosed in the present invention, first, multiple parameters of multiple processes of the entire process of the project to be evaluated are obtained, wherein the multiple parameters of each process include the actual completion period, the expected period, the actual investment and the expected investment. It should be understood that the actual completion period and the expected period of the process can provide time information about the completion of the process, which is used as a benchmark for evaluating the time consumption of rectification. The actual investment and the expected investment of the process can provide information about the resources required for the process, including manpower, materials and equipment, which have a certain impact on the evaluation of the time consumption of rectification.

[0026] Next, considering that there is a dependent relationship between each process in the whole process of the project to be evaluated, it will affect the transmission and scope of the rectification problem, thereby affecting the time consumption of the rectification. In addition, there is also a correlation between the multiple parameters of each process, which jointly affects the evaluation accuracy of the rectification time consumption. Therefore, in order to be able to capture and characterize the correlation characteristics between the multiple parameters of the multiple processes in the whole process of the project to be evaluated, in the technical solution disclosed in the present invention, the multiple parameters of the multiple processes are further arranged into a full-process multi-parameter matrix according to the parameter sample dimension and the process sample dimension. In this way, the distribution information of the multiple parameters of the multiple processes in the whole process of the project to be evaluated in the parameter sample dimension and the process sample dimension are respectively integrated.

[0027] Then, a process parameter correlation feature extractor based on a convolutional neural network model with excellent performance in extracting implicit correlation features is used to perform feature mining on the full process multi-parameter matrix, so as to extract the implicit correlation feature distribution information between each process and each parameter of the entire process of the project to be evaluated, thereby obtaining the full process parameter correlation feature vector.

[0028] Furthermore, for the problems that need to be rectified, first, obtain the text description of the problems that need to be rectified of the project to be evaluated. Next, considering that there is a large amount of semantic information in the text description of the problems that need to be rectified of the project to be evaluated, and the text description of the problems that need to be rectified of the project to be evaluated is composed of multiple words, the semantic association between each word expresses the overall semantics of the text description of the problems that need to be rectified of the project to be evaluated. Therefore, in the technical solution of the present disclosure, the text description of the problems that need to be rectified of the project to be evaluated is further processed by word segmentation to obtain a sequence of problem text description words, so as to divide the text description of the problems that need to be rectified of the project to be evaluated into meaningful words, which is helpful to further analyze and understand the theme, keywords and semantics of the problems that need to be rectified, thereby improving the accuracy of the subsequent time-consuming evaluation of the rectification problems.

[0029] Then, in order to be able to semantically understand the text description of the problems to be rectified in the project to be evaluated, it is also necessary to convert each word in the text description of the problems to be rectified in the project to be evaluated into a high-dimensional word vector representation, so as to capture the semantic association feature information between words later. Therefore, in the technical solution of the present disclosure, the sequence of the problem text description words is further extracted through the AlBert model of the semantic encoder to obtain a sequence of problem text description word vectors.

[0030] Furthermore, the sequence of the problem text description word vectors is semantically encoded in the text convolutional neural network model of the semantic encoder to extract the semantic association feature information between the words in the text description of the problems that need to be rectified in the project to be evaluated, thereby obtaining the problem text semantic understanding feature vector. It should be understood that the text convolutional neural network model is a model that can capture local features in the text. It extracts features from the text through convolution operations and can capture local semantic association feature information at different positions in the text description of the problems that need to be rectified in the project to be evaluated. Therefore, through the encoding of the text convolutional neural network model, the local semantics of the text description of the problems that need to be rectified in the project to be evaluated can be effectively captured.

[0031] Furthermore, the feature-to-feature attention layer is used to perform feature interaction based on the attention mechanism on the associated feature vector of the whole engineering process parameters and the semantic understanding feature vector of the problem text to obtain the semantic interaction feature vector of the engineering status-needs-rectification problem, so as to capture the association and mutual influence between the associated features of the whole engineering process parameters and the semantic features of the rectification problem. It should be understood that since the goal of the traditional attention mechanism is to learn an attention weight matrix, a larger weight is assigned to important features and a smaller weight is assigned to secondary features, so as to select more critical information for the current task goal. This method focuses more on weighting the importance of each feature and ignores the dependency between features. The feature-to-feature attention layer can capture the correlation and mutual influence between the associated features of the whole engineering process parameters and the semantic features of the rectification problem through feature interaction based on the attention mechanism, learn the dependency between different features, and interact and integrate the features according to these dependencies, so as to obtain the semantic interaction feature vector of the engineering status-needs-rectification problem.

[0032] Accordingly, if Figure 3 As shown, feature interaction analysis is performed on multiple parameters of the multiple processes and the text description of the problems to be rectified of the project to be evaluated to obtain semantic interaction features between the project status and the problems to be rectified, including: S131, extracting all-process parameter association features on the multiple parameters of the multiple processes to obtain a all-process parameter association feature vector; S132, semantically encoding the text description of the problems to be rectified of the project to be evaluated to obtain a problem text semantic understanding feature vector; and, S133, performing feature interaction on the all-process parameter association feature vector and the problem text semantic understanding feature vector to obtain a project status-problem to be rectified semantic interaction feature vector as the project status-problem to be rectified semantic interaction feature.

[0033] Accordingly, in step S131, Figure 4As shown, extracting the correlation features of all process parameters from the multiple parameters of the multiple processes to obtain the correlation feature vector of all process parameters includes: S1311, arranging the multiple parameters of the multiple processes into a multi-parameter matrix of all processes according to the parameter sample dimension and the process sample dimension; and, S1312, passing the multi-parameter matrix of all processes through the correlation feature extractor between process parameters based on the convolutional neural network model to obtain the correlation feature vector of all process parameters. It is worth mentioning that the convolutional neural network (CNN) is a deep learning model mainly used to process data with a grid structure. The correlation feature extractor between process parameters based on the convolutional neural network model is used to extract the correlation features between all process parameters. Specifically, the multiple parameters of the multiple processes are arranged into a multi-parameter matrix of all processes according to the parameter sample dimension and the process sample dimension, and then processed through the convolutional neural network model to extract the correlation features between all process parameters. The purpose of this is to capture the patterns and relationships between all process parameters so as to better understand and utilize these correlation features. These correlation features can be used for further analysis, prediction or decision-making, such as process optimization, anomaly detection, etc. The process parameter correlation feature extractor based on the convolutional neural network model can help extract the correlation features between all process parameters, thereby improving the understanding and application capabilities of process data.

[0034] Accordingly, in step S132, Figure 5As shown, the text description of the problem to be rectified of the project to be evaluated is semantically encoded to obtain a semantic understanding feature vector of the problem text, including: S1321, the text description of the problem to be rectified of the project to be evaluated is word segmented to obtain a sequence of problem text description words; and, S1322, the sequence of problem text description words is passed through a semantic encoder including an ALbert model and a text convolutional neural network model to obtain the semantic understanding feature vector of the problem text. It is worth mentioning that the ALbert model is a pre-trained language model based on the Transformer architecture, which has achieved good results in semantic understanding and representation learning tasks. The ALbert model adopts parameter sharing and sentence-level independent training strategies to reduce the number of model parameters and training time, while improving the performance of the model. The text convolutional neural network model is a model based on a convolutional neural network for processing text data. It uses convolution operations to capture local features in text and integrates these features through pooling operations. In step S1321, the text description of the problem to be rectified of the project to be evaluated is segmented and converted into a sequence of problem text description words. Then, in step S1322, a semantic encoder including an ALbert model and a text convolutional neural network model is used to process the sequence of problem text description words to obtain a semantic understanding feature vector of the problem text. The combination of the ALbert model and the text convolutional neural network model can help extract semantic information and features in the text and encode them into a feature vector of a fixed dimension. The purpose of this is to convert natural language text into a representation that can be understood and processed by a machine, so as to facilitate subsequent text analysis, classification, clustering and other tasks. Through the semantic understanding feature vector, the problem to be rectified of the project to be evaluated can be better understood and utilized, and further problem analysis and decision-making can be carried out.

[0035] Correspondingly, in step S133, the feature interaction between the full process parameter association feature vector and the problem text semantic understanding feature vector is performed to obtain the engineering state-needs rectification problem semantic interaction feature vector as the engineering state-needs rectification problem semantic interaction feature, including: using the feature inter-attention layer to perform feature interaction based on the attention mechanism on the full process parameter association feature vector and the problem text semantic understanding feature vector to obtain the engineering state-needs rectification problem semantic interaction feature vector. It should be understood that the main function of the feature interaction based on the attention mechanism is to capture the correlation between the full process parameter association feature vector and the problem text semantic understanding feature vector. By calculating the attention weight, the importance between different features can be determined, so as to better integrate these features. This interaction can help the model better understand the semantic relationship between the engineering state and the problem that needs rectification, and further improve the understanding and expression ability of the engineering state and the problem that needs rectification. Through the feature interaction based on the attention mechanism, the full process parameter association feature and the problem text semantic understanding feature can be effectively integrated to generate the engineering state-needs rectification problem semantic interaction feature. These interactive features can provide richer and more accurate information for subsequent engineering state assessment, rectification problem analysis and decision-making tasks. In other words, feature interaction based on the attention mechanism helps improve the model's ability to model the semantic relationship between engineering status and issues that need to be rectified.

[0036] Next, the semantic interaction feature vector between the engineering status and the problem to be rectified is passed through a decoder to obtain a decoded value, which is used to represent the estimated duration for the problem to be rectified. In other words, decoding is performed based on the interactive correlation feature information between the correlation features of the entire engineering process parameters and the semantic features of the rectification problem, thereby determining the estimated duration of the problem to be rectified. In this way, the subjectivity and inefficiency problems caused by manual evaluation can be avoided, and the mutual influence of each process link and each factor as well as the semantic information of the problem to be rectified can be used to provide a more accurate estimation of the rectification time consumption, providing strong support for engineering management.

[0037] Accordingly, based on the semantic interaction features between the engineering status and the problem to be rectified, the estimated duration for the problem to be rectified is determined, including: passing the semantic interaction feature vector between the engineering status and the problem to be rectified through a decoder to obtain a decoded value, and the decoded value is used to represent the estimated duration for the problem to be rectified. It should be understood that the decoder is a part of the neural network model, which is used to convert the input feature vector into a specific output value or probability distribution. In the present disclosure, the role of the decoder is to convert the semantic interaction feature vector between the engineering status and the problem to be rectified into a decoded value representing the estimated duration for the problem to be rectified. The decoder is usually composed of a series of neural network layers, which can be a fully connected layer, a recurrent neural network (such as LSTM or GRU), etc., which is responsible for further processing and conversion of the input features to generate the final output result. The decoder receives the semantic interaction feature vector between the engineering status and the problem to be rectified as input, and maps it to a decoded value representing the estimated duration for the problem to be rectified through learning. The training process of the decoder will learn how to extract information related to the estimated duration from the feature vector based on the existing labeled data, and generate a corresponding decoded value. Through the decoder, the semantic interaction features of the project status and the problems that need to be corrected can be converted into specific estimated durations, thereby providing quantitative results for the time assessment of the problems that need to be corrected. Such estimated durations can be used for decision-making and planning in aspects such as project management, schedule control, and resource allocation.

[0038] Further, in an embodiment of the present disclosure, the engineering consulting service evaluation method further includes a training step: for training the process parameter correlation feature extractor based on the convolutional neural network model, the semantic encoder including the ALbert model and the text convolutional neural network model, and the decoder. It should be understood that the training step plays a vital role in the engineering consulting service evaluation method. Through training, the model can learn to extract useful features from the input data and convert it into a decoded value of the estimated time length of the problem to be rectified. In the training process of the process parameter correlation feature extractor, the model learns how to extract correlation features from all process parameters. These correlation features can capture the interaction and dependency between process parameters and provide useful input for subsequent feature interactions. By training the feature extractor, the model can better understand and represent the engineering status. In the training process of the semantic encoder, the model learns how to convert the text description of the problem to be rectified of the project to be evaluated into a machine-understandable feature vector. The ALbert model and the text convolutional neural network model are used as semantic encoders, which can capture the semantic information and context association of the text. By training the semantic encoder, the model can better understand and represent the semantics of the problem to be rectified. During the training process of the decoder, the model learns how to convert the semantic interaction feature vector between the project status and the problem to be corrected into a decoded value representing the estimated duration. By training the decoder, the model can learn how to extract information related to the estimated duration from the interaction features. The training process of the decoder can be supervised based on the existing labeled data to minimize the error between the predicted value and the true value. Through these training steps, the model can gradually learn how to effectively extract features, understand semantics, perform feature interactions, and generate estimated durations for problems that need to be corrected. Such a training process can improve the prediction accuracy and generalization ability of the model, making it more reliable and effective in practical applications.

[0039] More specifically, if Figure 6As shown, the training step includes: S210, obtaining training data, the training data including multiple training parameters of multiple processes of the whole process of the project to be evaluated, the training text description of the problems to be rectified of the project to be evaluated, and the true value of the estimated duration of the problems to be rectified; S220, arranging the multiple training parameters of the multiple processes into a training full-process multi-parameter matrix according to the parameter sample dimension and the process sample dimension; S230, passing the training full-process multi-parameter matrix through the process parameter correlation feature extractor based on the convolutional neural network model to obtain a training full-process parameter correlation feature vector; S240, performing word segmentation processing on the training text description of the problems to be rectified of the project to be evaluated to obtain a sequence of training problem text description words; S250, passing the sequence of training problem text description words through the ALbert model and the text convolutional neural network model A semantic encoder is used to obtain a semantic understanding feature vector of the training problem text; S260, using an inter-feature attention layer to perform feature interaction based on an attention mechanism on the training full-project process parameter association feature vector and the training problem text semantic understanding feature vector to obtain a training project status-issue-needs-correction semantic interaction feature vector; S270, passing the training project status-issue-needs-correction semantic interaction feature vector through the decoder to obtain a decoding loss function value; and, S280, based on the decoding loss function value and through back propagation of gradient descent, the process parameter association feature extractor based on the convolutional neural network model, the semantic encoder including the ALbert model and the text convolutional neural network model, and the decoder are trained, wherein, in each round of iteration of the training process, the training project status-issue-needs-correction semantic interaction feature vector is optimized.

[0040] In particular, in the technical solution of the present application, when the training full-process multi-parameter matrix is ​​obtained by the process parameter correlation feature extractor based on the convolutional neural network model to obtain the training full-process parameter correlation feature vector, each eigenvalue of the training full-process parameter correlation feature vector is used to represent the local correlation value of the parameter under the sample-process cross dimension, and each eigenvalue of the training problem text semantic understanding feature vector expresses the semantic features of the sequence of the problem text description words based on the ALbert model, and the features of the local correlation semantic feature extraction are performed via the text convolutional neural network model. Therefore, after using the inter-feature attention layer to perform feature interaction based on the attention mechanism on the training full-project process parameter correlation feature vector and the training problem text semantic understanding feature vector, the obtained training engineering status-needs-rectification problem semantic interaction feature vector will also have the feature value granularity super-resolution expression characteristics based on the multi-dimensional context of the dimensions of the full-project process parameter correlation feature vector and the problem text semantic understanding feature vector, so that the convergence mapping iteration consistency of the training engineering status-needs-rectification problem semantic interaction feature vector to the predetermined probability category representation in the decoding probability space is poor, thereby reducing the accuracy of the decoding result.

[0041] Based on this, the present application optimizes the interaction feature vector between the training project status and the semantics of the problem to be rectified in each iteration of the training process, including: the first eigenvalue of the interaction feature vector between the training project status and the semantics of the problem to be rectified and the second eigenvalue , calculate the absolute value of the difference divided by the sum of the absolute values ​​and then correct the relative probability value, wherein the correction makes the relative probability value greater than zero and less than one; calculate the first product of the first eigenvalue multiplied by the relative probability value, the second product of the second eigenvalue multiplied by the absolute value of the difference between the relative probability value and one, the first quotient of the first eigenvalue divided by the absolute value of the difference between the relative probability value and one, and the second quotient of the second eigenvalue divided by the relative probability value; after adding the first product and the second product, further subtract the difference between the first quotient and the second quotient to obtain the value equal to the first eigenvalue and the second eigenvalue The corresponding eigenvalues ​​of the corresponding optimization matrix ; and, multiplying the optimization matrix by the training engineering state-issue semantic interaction feature vector to obtain an optimized training engineering state-issue semantic interaction feature vector.

[0042] Specifically, it is expressed as: ;in, Represents the semantic interaction feature vector between the training project status and the problem to be rectified, and respectively represent the first eigenvalue and the second eigenvalue of the interaction feature vector between the training project status and the semantics of the problem to be rectified, represents the relative probability value, Represents the first eigenvalue and the second eigenvalue The corresponding eigenvalues ​​of the corresponding optimization matrix, Represents the first eigenvalue and the second eigenvalue The corresponding optimization matrix is, represents matrix multiplication, Represents the interaction feature vector between the optimized training project status and the semantics of the problem that needs to be corrected.

[0043] Among them, the first eigenvalue and the second eigenvalue The corresponding relative probability value It is expressed as: ;in, and respectively represent the first eigenvalue and the second eigenvalue of the interaction feature vector between the training project status and the semantics of the problem to be rectified, represents the predetermined hyperparameters, Represents the relative probability value.

[0044] Here, for the region-boundary integral relationship of the high-dimensional feature manifold of the interaction feature vector between the training engineering state and the semantics of the problem to be rectified, the probability distribution of the feature set of the training engineering state-the semantics of the interaction feature vector between the training engineering state and the problem to be rectified is defined as the manifold constraint boundary to approximate the high-dimensional manifold of the feature set of the training engineering state-the semantics of the interaction feature vector between the training engineering state and the problem to be rectified, which is composed of eigenvalue pairs. The single-connected region representation is composed of eigenvalue pairs, so as to avoid the mapping ambiguity of the diversified feature representation of the interaction feature vector between the training engineering state-the semantics of the problem to be rectified to the local manifold representation in the high-dimensional feature decoding regression convergence space, and improve the execution iteration consistency of each local feature distribution in the mapping task, so as to improve the feature decoding regression convergence effect and the accuracy of the decoding result. In this way, the time consumption of the rectification problem can be automatically evaluated based on the correlation relationship between multiple parameters of multiple processes in the whole process of the project and the semantic information of the problem to be rectified, so as to provide a more accurate estimation of the rectification time consumption and provide strong support for project management.

[0045] In summary, the engineering consulting service evaluation method based on the embodiment of the present disclosure can avoid the subjectivity and inefficiency problems brought about by manual evaluation, and also utilizes the mutual influence and correlation between various process links and various factors to provide a more accurate estimate of the time consumption for rectification, thereby providing strong support for engineering management.

[0046] Figure 7 FIG. 1 is a block diagram of an engineering consulting service evaluation system 100 according to an embodiment of the present disclosure. Figure 7 As shown, the engineering consulting service evaluation system 100 according to the embodiment of the present disclosure includes: a parameter acquisition module 110, which is used to obtain multiple parameters of multiple processes in the whole process of the project to be evaluated, wherein the multiple parameters of each process include the actual completion period, the expected period, the actual investment and the expected investment; a problem acquisition module 120, which is used to obtain a text description of the problem to be rectified of the project to be evaluated; a feature interaction analysis module 130, which is used to perform feature interaction analysis on the multiple parameters of the multiple processes and the text description of the problem to be rectified of the project to be evaluated to obtain the interaction features between the project status and the problem to be rectified; and a duration estimation module 140, which is used to determine the estimated duration for the problem to be rectified based on the interaction features between the project status and the problem to be rectified.

[0047] Here, those skilled in the art will appreciate that the specific functions and operations of the various units and modules in the above-mentioned engineering consulting service evaluation system 100 have been described in detail above. Figures 1 to 6 The description of the engineering consulting service evaluation method has been introduced in detail, and therefore, its repeated description will be omitted.

[0048] As described above, the engineering consulting service evaluation system 100 according to the embodiment of the present disclosure can be implemented in various wireless terminals, such as a server with an engineering consulting service evaluation algorithm. In a possible implementation, the engineering consulting service evaluation system 100 according to the embodiment of the present disclosure can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the engineering consulting service evaluation system 100 can be a software module in the operating system of the wireless terminal, or can be an application developed for the wireless terminal; of course, the engineering consulting service evaluation system 100 can also be one of the many hardware modules of the wireless terminal.

[0049] Alternatively, in another example, the engineering consulting service evaluation system 100 and the wireless terminal may also be separate devices, and the engineering consulting service evaluation system 100 may be connected to the wireless terminal via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0050] Figure 8 FIG. 1 shows an application scenario diagram of the engineering consulting service evaluation method according to an embodiment of the present disclosure. Figure 8 As shown, in this application scenario, first, multiple parameters of multiple processes of the entire process of the project to be evaluated are obtained (for example, Figure 8 D1 as indicated in the figure), and a textual description of the issues that need to be corrected for the project to be assessed (e.g. Figure 8D2 as shown in the figure), wherein the multiple parameters of each process include the actual completion period, the expected period, the actual investment and the expected investment, and then the multiple parameters of the multiple processes and the text description of the problems to be rectified of the project to be evaluated are input into a server deployed with an engineering consulting service evaluation algorithm (for example, Figure 8 S), wherein the server is capable of using the engineering consulting service evaluation algorithm to process multiple parameters of the multiple processes and a text description of the problems to be corrected of the project to be evaluated to obtain a decoded value representing an estimated duration for the problems to be corrected.

[0051] The flow chart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to multiple embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a part of a module, program segment or instruction, and a part of the module, program segment or instruction includes one or more executable instructions for realizing the specified logical function. In some alternative implementations, the function marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous square boxes can actually be executed substantially in parallel, and they can sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or action, or can be implemented with a combination of special hardware and computer instructions.

[0052] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A method for evaluating engineering consulting services, characterized in that: include: Acquire multiple parameters of multiple processes of the entire process of the project to be evaluated, wherein the multiple parameters of each process include the actual completion period, the expected period, the actual investment and the expected investment; obtain a text description of the problems to be rectified of the project to be evaluated; perform feature interaction analysis on the multiple parameters of the multiple processes and the text description of the problems to be rectified of the project to be evaluated to obtain semantic interaction features between the project status and the problems to be rectified; and determine the estimated duration for the problems to be rectified based on the semantic interaction features between the project status and the problems to be rectified; Wherein, based on the semantic interaction feature between the engineering status and the problem to be rectified, determining the estimated duration for the problem to be rectified includes: passing the semantic interaction feature vector between the engineering status and the problem to be rectified through a decoder to obtain a decoded value, wherein the decoded value is used to represent the estimated duration for the problem to be rectified; The training step is also included: for training a process parameter correlation feature extractor based on a convolutional neural network model, a semantic encoder and a decoder including an ALbert model and a text convolutional neural network model; The training step includes: obtaining training data, wherein the training data includes multiple training parameters of multiple processes of the entire process of the project to be evaluated, a training text description of the problems to be rectified of the project to be evaluated, and the true value of the estimated duration of the problems to be rectified; arranging the multiple training parameters of the multiple processes into a training full-process multi-parameter matrix according to the parameter sample dimension and the process sample dimension; passing the training full-process multi-parameter matrix through the process parameter correlation feature extractor based on the convolutional neural network model to obtain a training full-process parameter correlation feature vector; performing word segmentation processing on the training text description of the problems to be rectified of the project to be evaluated to obtain a sequence of training problem text description words; passing the sequence of training problem text description words through the semantic encoding method including the ALbert model and the text convolutional neural network model The invention relates to a method for extracting process parameters from a convolutional neural network model and a semantic encoder comprising an ALbert model and a text convolutional neural network model, wherein the semantic encoder comprises an ALbert model and a text convolutional neural network model, and the decoder. The method comprises the following steps: using an inter-feature attention layer to perform feature interaction based on an attention mechanism on the training full process parameter association feature vector and the training problem text semantic understanding feature vector to obtain a training engineering state-need-to-correction problem semantic interaction feature vector; passing the training engineering state-need-to-correction problem semantic interaction feature vector through the decoder to obtain a decoding loss function value; and training the process parameter association feature extractor based on the convolutional neural network model, the semantic encoder comprising an ALbert model and a text convolutional neural network model, and the decoder based on the decoding loss function value and through back propagation of gradient descent, wherein the training engineering state-need-to-correction problem semantic interaction feature vector is optimized in each iteration of the training process; In each iteration of the training process, the interaction feature vector between the training project status and the semantics of the problem to be rectified is optimized, including: the first eigenvalue of the interaction feature vector between the training project status and the semantics of the problem to be rectified and the second eigenvalue , calculate the absolute value of the difference divided by the sum of the absolute values ​​and then correct the relative probability value, wherein the correction makes the relative probability value greater than zero and less than one; calculate the first product of the first eigenvalue multiplied by the relative probability value, the second product of the second eigenvalue multiplied by the absolute value of the difference between the relative probability value and one, the first quotient of the first eigenvalue divided by the absolute value of the difference between the relative probability value and one, and the second quotient of the second eigenvalue divided by the relative probability value; after adding the first product and the second product, further subtract the difference between the first quotient and the second quotient to obtain the value equal to the first eigenvalue and the second eigenvalue The corresponding eigenvalues ​​of the corresponding optimization matrix ; and, multiplying the optimization matrix by the training engineering state-issue semantic interaction feature vector to obtain an optimized training engineering state-issue semantic interaction feature vector.

2. The engineering consulting service evaluation method according to claim 1, characterized in that: A feature interaction analysis is performed on the multiple parameters of the multiple processes and the text description of the problems to be rectified of the project to be evaluated to obtain semantic interaction features between the project status and the problems to be rectified, including: extracting all-process parameter association features on the multiple parameters of the multiple processes to obtain a all-process parameter association feature vector; semantically encoding the text description of the problems to be rectified of the project to be evaluated to obtain a problem text semantic understanding feature vector; and feature interaction is performed on the all-process parameter association feature vector and the problem text semantic understanding feature vector to obtain a project status-problem to be rectified semantic interaction feature vector as the project status-problem to be rectified semantic interaction feature.

3. The engineering consulting service evaluation method according to claim 2, characterized in that: The full-process parameter association feature extraction is performed on the multiple parameters of the multiple processes to obtain the full-process parameter association feature vector, including: arranging the multiple parameters of the multiple processes into a full-process multi-parameter matrix according to the parameter sample dimension and the process sample dimension; and passing the full-process multi-parameter matrix through a process parameter association feature extractor based on a convolutional neural network model to obtain the full-process parameter association feature vector.

4. The engineering consulting service evaluation method according to claim 3 is characterized in that: The text description of the problem to be rectified of the project to be evaluated is semantically encoded to obtain a semantic understanding feature vector of the problem text, including: performing word segmentation processing on the text description of the problem to be rectified of the project to be evaluated to obtain a sequence of problem text description words; and passing the sequence of problem text description words through a semantic encoder including an ALbert model and a text convolutional neural network model to obtain the semantic understanding feature vector of the problem text.

5. The engineering consulting service evaluation method according to claim 4, characterized in that: Performing feature interaction on the feature vector associated with all process parameters and the feature vector for semantic understanding of the problem text to obtain a semantic interaction feature vector of the project status-issues requiring rectification as the semantic interaction feature of the project status-issues requiring rectification, including: using an inter-feature attention layer to perform feature interaction based on an attention mechanism on the feature vector associated with all process parameters and the feature vector for semantic understanding of the problem text to obtain the semantic interaction feature vector of the project status-issues requiring rectification.

6. An engineering consulting service evaluation system according to the engineering consulting service evaluation method according to claim 1, characterized in that: include: A parameter acquisition module is used to acquire multiple parameters of multiple processes in the entire process of the project to be evaluated, wherein the multiple parameters of each process include the actual completion period, the expected period, the actual investment and the expected investment; a problem acquisition module is used to acquire a text description of the problem to be rectified of the project to be evaluated; a feature interaction analysis module is used to perform feature interaction analysis on the multiple parameters of the multiple processes and the text description of the problem to be rectified of the project to be evaluated to obtain the interaction features between the project status and the problem to be rectified semantics; and a duration estimation module, which is used to determine the estimated duration for the problem to be rectified based on the semantic interaction characteristics between the project status and the problem to be rectified.

Images