Project status prediction system based on project status feature changes

Abstract

The present application relates to the field of intelligent management, and specifically discloses a project status prediction system based on changes in project status characteristics, which extracts the time series of multidimensional features of the monitored project from the background database, and uses data processing and analysis technology based on deep learning neural networks to extract and associate the project features of each dimension in the time series of the multidimensional features of the project, so as to intelligently obtain the monitoring results of whether there are abnormalities in the project progress based on the multidimensional time series aggregation features of each project. Through this method, a more comprehensive status analysis can be provided to timely update the project progress, and data-driven decision support can be provided for the project management team to formulate more effective management strategies.

Description

Technical Field

[0001] The present application relates to the field of intelligent management, and more specifically, to a project status prediction system based on changes in project status characteristics. Background Art

[0002] As the complexity of project management increases, there is a growing need for accurate predictions of project status to ensure that projects are completed on time, on budget and to quality requirements.

[0003] Project status forecasting is a management practice that involves the use of various methods and tools to assess and predict the status of a project at a specific point in time. This forecast helps the project team, stakeholders, and management understand whether the project progress is in line with expectations and make corresponding adjustments and decisions.

[0004] However, traditional project status prediction methods often lack the ability to model the complex relationships between project features, making it difficult to capture all key factors, resulting in low accuracy in prediction results. In addition, especially for time series data, traditional methods often require manual feature extraction or use simple statistical methods for modeling, which cannot fully utilize the time correlation and dynamic information in the data for prediction, resulting in low credibility of prediction results.

[0005] Therefore, a project status prediction system based on changes in project status characteristics is desired. Summary of the invention

[0006] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a project status prediction system based on the change of project status characteristics, which extracts the time series of the project multidimensional characteristics of the monitored project from the background database, and uses the data processing and analysis technology based on the deep learning neural network to extract and associate the project characteristics of each dimension in the time series of the project multidimensional characteristics, so as to intelligently obtain the monitoring results of whether there are abnormalities in the project progress according to the multidimensional time series aggregation characteristics of each project. Through this method, a more comprehensive status analysis can be provided to update the project progress in a timely manner, and data-driven decision support can be provided to the project management team to formulate more effective management strategies.

[0007] According to one aspect of the present application, a project status prediction system based on changes in project status characteristics is provided, comprising:

[0008] A project feature acquisition module is used to extract the time series of multi-dimensional features of the monitored project from the background database;

[0009] A data regularization module, used for performing data regularization on the time series of the multi-dimensional features of the project to obtain time series of multiple groups of project features;

[0010] A feature time series aggregation module, used to extract multiple project feature time series aggregation representation vectors from the time series of the multiple groups of project features;

[0011] A feature aggregation module, used for performing feature aggregation on the plurality of project feature time series aggregation representation vectors to obtain a project multi-dimensional feature time series aggregation representation vector as a project multi-dimensional feature time series aggregation representation feature;

[0012] A monitoring result generation module, used for obtaining monitoring results based on the time series aggregation representation features of the multi-dimensional features of the project;

[0013] Wherein, the feature time series aggregation module includes:

[0014] A project feature embedding coding unit, for embedding and coding each project feature in the time series of each group of project features in the time series of the multiple groups of project features to obtain a time series of project feature embedding coding vectors;

[0015] A time series association implicit pattern feature extraction unit, used for extracting the time series association implicit pattern feature of the project features from the time series of the project feature embedding coding vector to obtain a sequence of project feature time series association implicit representation vectors;

[0016] The information transfer aggregation unit is used to pass the sequence of the project feature time series associated implicit representation vectors through the information transfer aggregation network to obtain the project feature time series aggregation representation vector.

[0017] In the above-mentioned project status prediction system based on changes in project status features, the data regularization module is used to: based on feature dimensions, perform data regularization on the time series of the multi-dimensional features of the project to obtain the time series of the multiple groups of project features.

[0018] In the above-mentioned project status prediction system based on changes in project status features, the time series association implicit pattern feature extraction unit is used to: embed the project features into the time series of the encoding vector through a project feature time series association implicit pattern feature extractor based on a bidirectional LSTM model to obtain a sequence of the project feature time series association implicit representation vectors.

[0019] In the above-mentioned project status prediction system based on changes in project status features, the information transmission aggregation unit includes: a temporal association implicit aggregation feature calculation subunit, which is used to calculate the project feature temporal association implicit aggregation feature vectors corresponding to each project feature temporal association implicit representation vector in the sequence of the project feature temporal association implicit representation vectors; and an aggregation subunit, which is used to positionally sum the project feature temporal association implicit aggregation feature vectors corresponding to each project feature temporal association implicit representation vector in the sequence of the project feature temporal association implicit representation vectors to obtain the project feature temporal association representation vector.

[0020] In the above-mentioned project status prediction system based on project status feature changes, the temporal correlation implicit aggregation feature calculation subunit includes: a feature summation secondary subunit for calculating the first in the sequence of the project feature temporal correlation implicit representation vector The weighted sum of the implicit representation vectors of the temporal association of the project features is obtained to obtain the implicit sum representation vector of the temporal association of the project features; the feature weighted secondary subunit is used to use the sum of the trainable preset hyperparameter and one as the weighting coefficient to The weighted processing is performed on each position feature value in the item feature time series association implicit representation vector to obtain a weighted item feature time series association implicit representation vector; and a multi-layer perception secondary subunit is used to sum the item feature time series association implicit sum representation vector and the weighted item feature time series association implicit representation vector by position and then process them through a multi-layer perception machine to obtain a first The project feature time series association implicit aggregation feature vector corresponding to the project feature time series association implicit representation vector.

[0021] In the above-mentioned project status prediction system based on changes in project status features, the feature aggregation module is used to: use a feature aggregator based on a multi-layer perceptron model to process the multiple project feature time series aggregation representation vectors to obtain the project multi-dimensional feature time series aggregation representation vector.

[0022] In the above-mentioned project status prediction system based on changes in project status characteristics, the monitoring result generation module is used to: pass the project multi-dimensional feature time series aggregation representation vector through a classifier-based project progress intelligent monitor to obtain a monitoring result, and the monitoring result is used to indicate whether there is an abnormality in the project progress.

[0023] Compared with the prior art, the project status prediction system provided by the present application is based on the change of project status characteristics. It extracts the time series of the multidimensional characteristics of the project being monitored from the background database, and uses the data processing and analysis technology based on the deep learning neural network to extract and associate the project characteristics of each dimension in the time series of the multidimensional characteristics of the project, so as to intelligently obtain the monitoring results of whether there are abnormalities in the project progress according to the multidimensional time series aggregation characteristics of each project. Through this method, a more comprehensive status analysis can be provided to update the project progress in a timely manner, and data-driven decision support can be provided to the project management team to formulate more effective management strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0025] Figure 1 It is a block diagram of a project status prediction system based on changes in project status characteristics according to an embodiment of the present application.

[0026] Figure 2 Schematic diagram of the architecture of a project status prediction system based on changes in project status characteristics according to an embodiment of the present application.

[0027] Figure 3 It is a block diagram of a feature time series aggregation module in a project status prediction system based on changes in project status features according to an embodiment of the present application.

[0028] Figure 4 It is a block diagram of an information transmission aggregation unit in a project status prediction system based on changes in project status characteristics according to an embodiment of the present application.

[0029] Figure 5 It is a block diagram of a temporal correlation implicit aggregation feature calculation subunit in a project status prediction system based on changes in project status features according to an embodiment of the present application. DETAILED DESCRIPTION

[0030] The embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, but rather these embodiments are provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0031] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.

[0032] In the description of the embodiments of the present disclosure, the term "including" and similar terms should be understood as open inclusion, that is, "including but not limited to". The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc. may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0033] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0034] As the complexity of project management increases, there is a growing need for accurate predictions of project status to ensure that projects are completed on time, on budget and to quality requirements.

[0035] Project status forecasting is a management practice that involves the use of various methods and tools to assess and predict the status of a project at a specific point in time. This forecast helps the project team, stakeholders, and management understand whether the project progress is in line with expectations and make corresponding adjustments and decisions.

[0036] However, traditional project status prediction methods often lack the ability to model the complex relationships between project features, making it difficult to capture all key factors, resulting in low accuracy in prediction results. In addition, especially for time series data, traditional methods often require manual feature extraction or use simple statistical methods for modeling, which cannot fully utilize the time correlation and dynamic information in the data for prediction, resulting in low credibility of prediction results.

[0037] Based on this, the present application proposes a project status prediction system based on changes in project status features, which extracts the time series of the project multidimensional features of the monitored project from the background database, and uses data processing and analysis technology based on deep learning neural networks to extract and associate the project features of each dimension in the time series of the project multidimensional features, so as to intelligently obtain the monitoring results of whether there are abnormalities in the project progress based on the multidimensional time series aggregation features of each project. Through this method, a more comprehensive status analysis can be provided to update the project progress in a timely manner, and data-driven decision support can be provided to the project management team to formulate more effective management strategies.

[0038] Figure 1 It is a block diagram of a project status prediction system based on changes in project status characteristics according to an embodiment of the present application. Figure 2 FIG. 1 is a schematic diagram of the architecture of a project status prediction system based on changes in project status characteristics according to an embodiment of the present application. Figure 1 and Figure 2 As shown, according to the project status prediction system 100 based on the change of project status characteristics according to the embodiment of the present application, it includes: a project feature acquisition module 110, which is used to extract the time series of project multidimensional characteristics of the monitored project from the background database; a data regularization module 120, which is used to perform data regularization on the time series of the project multidimensional characteristics to obtain multiple groups of project feature time series; a feature time series aggregation module 130, which is used to extract multiple project feature time series aggregation representation vectors from the time series of the multiple groups of project features; a feature aggregation module 140, which is used to perform feature aggregation on the multiple project feature time series aggregation representation vectors to obtain a project multidimensional feature time series aggregation representation vector as a project multidimensional feature time series aggregation representation feature; and a monitoring result generation module 150, which is used to obtain a monitoring result based on the project multidimensional feature time series aggregation representation feature.

[0039] In an embodiment of the present application, the project feature acquisition module 110 is used to extract the time series of the project multidimensional features of the monitored project from the background database. It should be understood that the project multidimensional features of the monitored project refer to a series of key indicators and attributes that can reflect the project status in the project management process. Specifically, the project multidimensional features may include time-related features, cost-related features, progress-related features, etc. These multidimensional feature information plays an important role and influence on the evaluation of the project. Based on this, in order to more accurately and comprehensively evaluate the progress of the project, in the technical solution of the present application, the time series of the project multidimensional features of the monitored project is extracted from the background database.

[0040] In an embodiment of the present application, the data regularization module 120 is used to perform data regularization on the time series of the multidimensional features of the project to obtain the time series of multiple groups of project features. Specifically, in an embodiment of the present application, the data regularization module is used to: based on the feature dimension, perform data regularization on the time series of the multidimensional features of the project to obtain the time series of the multiple groups of project features. Accordingly, considering that the multidimensional features of the project can be very diverse, for example, time-related features: project start and end dates, start and end dates of each stage, completion time of key milestones; cost-related features: total budget, amount spent, cost deviation; progress-related features: percentage of project completion, remaining workload estimate, progress deviation, etc. These different feature dimensions have different data distributions, dimensions and ranges of variation. Based on this, in order to analyze and process the feature information of each dimension in the time series of the multidimensional features of the project in a more fine-grained manner, so as to effectively extract the temporal features of each dimension and perform feature association, in the technical solution of the present application, based on the feature dimension, the time series of the multidimensional features of the project is data regularized to obtain the time series of multiple groups of project features. That is, by regularizing the time series of the multidimensional features of the project based on the feature dimension, the entire project multidimensional data can be segmented to obtain the time series of projects with multiple different feature dimensions, and the project information of each feature dimension can be individually extracted and correlated, so that the system can better capture the temporal relationship between project features of different dimensions, thereby helping the project status prediction system to more accurately and efficiently monitor whether there are any abnormalities in the project progress.

[0041] In an embodiment of the present application, the feature time series aggregation module 130 is used to extract multiple project feature time series aggregation representation vectors from the time series of the multiple groups of project features. It should be understood that each project feature in the time series of the multiple groups of project features involves multiple different contents, such as budget, cost, expenditure, etc. Therefore, in order to integrate the information of each content to form a comprehensive project feature representation to more comprehensively describe the status and characteristics of each project, in the technical solution of the present application, multiple project feature time series aggregation representation vectors are extracted from the time series of the multiple groups of project features, which can better reflect the overall status and feature information of the project features of each dimension, and provide a better basis and support for the subsequent feature analysis and evaluation of the comprehensive situation of the entire project, thereby helping project managers or regulatory authorities to better understand the overall progress of the project, so as to promptly discover potential problems and take measures to adjust them.

[0042] Figure 3 is a block diagram of a feature time series aggregation module in a project status prediction system based on project status feature changes according to an embodiment of the present application. Specifically, in the embodiment of the present application, Figure 3As shown, the feature time series aggregation module 130 includes: a project feature embedding coding unit 131, which is used to embed and code each project feature in the time series of each group of project features in the time series of the multiple groups of project features to obtain a time series of project feature embedding coding vectors; a time series association implicit pattern feature extraction unit 132, which is used to extract project feature time series association implicit pattern features from the time series of the project feature embedding coding vectors to obtain a sequence of project feature time series association implicit representation vectors; and an information transfer aggregation unit 133, which is used to pass the sequence of project feature time series association implicit representation vectors through an information transfer aggregation network to obtain the project feature time series aggregation representation vector.

[0043] Specifically, the project feature embedding coding unit 131 is used for embedding and coding each project feature in the time series of each group of project features in the time series of the multiple groups of project features to obtain a time series of project feature embedding coding vectors. Correspondingly, each project feature in the time series of project features belongs to high-dimensional project feature data. Therefore, in order to map these high-dimensional project feature data into a low-dimensional embedding space to facilitate subsequent feature extraction and capture, in the technical solution of the present application, each project feature in the time series of project features is embedded and coded to obtain a time series of project feature embedding coding vectors. Specifically, through embedded coding, the original project features can be converted into a higher-level abstract representation, that is, the original sparse project feature data can be converted into a dense vector representation, which helps to capture the potential patterns and trends of the project status, so that each project feature learns a more semantically rich representation, thereby obtaining a time series of project feature embedding coding vectors containing more complex patterns.

[0044] Specifically, the temporal association implicit pattern feature extraction unit 132 is used to extract the temporal association implicit pattern features of the project features from the time series of the project feature embedded coding vector to obtain a sequence of project feature temporal association implicit representation vectors. Specifically, in an embodiment of the present application, the temporal association implicit pattern feature extraction unit is used to: pass the time series of the project feature embedded coding vector through a project feature temporal association implicit pattern feature extractor based on a bidirectional LSTM model to obtain a sequence of project feature temporal association implicit representation vectors. It should be understood that, considering that in project management, project features often change over time, that is, there is a certain temporal correlation between the various project features. There is a mutual temporal correlation relationship and implicit pattern information between the various project feature embedded coding vectors in the time series of the project feature embedded coding vector. In addition, considering the bidirectional LSTM model, the temporal correlation between time series data can be effectively captured to help the model better understand the evolution laws and trends between time series features. Based on this, in the technical solution of the present application, the time series of the project feature embedding coding vector is passed through the project feature temporal association implicit pattern feature extractor based on the bidirectional LSTM model to obtain a sequence of project feature temporal association implicit representation vectors. Specifically, the bidirectional LSTM model combines the forward and reverse information flows, which can better capture the long-term dependencies in the time series of the project feature embedding coding vector, and help discover the important features and patterns of potential project features in the time series of the project feature embedding coding vector, so as to obtain a more representative project feature temporal association implicit representation, thereby improving the model's understanding and prediction capabilities of project status temporal changes.

[0045] Specifically, the information transfer aggregation unit 133 is used to pass the sequence of the project feature temporal association implicit representation vectors through the information transfer aggregation network to obtain the project feature temporal aggregation representation vector. Accordingly, it is considered that there are complex information association relationships and context dependencies between the individual project feature temporal association implicit representation vectors in the sequence of the project feature temporal association implicit representation vectors. Therefore, in order to help the system better understand and utilize the relationship between project features, so as to improve the overall grasp and representation capabilities of the state of each project feature dimension, in the technical solution of the present application, the sequence of the project feature temporal association implicit representation vectors is passed through the information transfer aggregation network to obtain the project feature temporal aggregation representation vector. It is worth mentioning that the information transfer aggregation network usually adopts the idea of ​​graph neural network, which can effectively process unstructured data and complex relational structures. Specifically, in project management, the relationship between project features can be viewed as a graph structure. The information transmission aggregation network can effectively model the graph structure relationship between the implicit representation features of the temporal association of each project feature, and capture the interactive information and important feature influences between the implicit features of the temporal association of each project feature by transmitting information in the network. This helps the system to more comprehensively understand the correlation between the features of each project, thereby improving the ability to model changes in project status.

[0046] Figure 4 is a block diagram of an information transmission aggregation unit in a project status prediction system based on project status feature changes according to an embodiment of the present application. Specifically, in the embodiment of the present application, Figure 4 As shown, the information transmission aggregation unit 133 includes: a temporal association implicit aggregation feature calculation subunit 1331, which is used to calculate the project feature temporal association implicit aggregation feature vector corresponding to each project feature temporal association implicit representation vector in the sequence of the project feature temporal association implicit representation vector; and an aggregation subunit 1332, which is used to positionally sum the project feature temporal association implicit aggregation feature vectors corresponding to each project feature temporal association implicit representation vector in the sequence of the project feature temporal association implicit representation vector to obtain the project feature temporal association representation vector.

[0047] Figure 5 is a block diagram of a time-series-related implicit aggregation feature calculation subunit in a project status prediction system based on project status feature changes according to an embodiment of the present application. More specifically, in the embodiment of the present application, Figure 5 As shown, the temporal correlation implicit aggregation feature calculation subunit 1331 includes: a feature summing secondary subunit 13311, which is used to calculate the first in the sequence of the project feature temporal correlation implicit representation vector. The weighted sum of the implicit representation vectors of the temporal association of project features is obtained to obtain the implicit sum representation vector of the temporal association of project features; the feature weighted secondary subunit 13312 is used to use the sum of the trainable preset hyperparameter and one as the weighting coefficient to weight the first The weighted processing is performed on each position feature value in the implicit representation vector of the project feature temporal association to obtain a weighted project feature temporal association implicit representation vector; and the multi-layer perception secondary subunit 13313 is used to sum the project feature temporal association implicit sum representation vector and the weighted project feature temporal association implicit representation vector by position and then process them through a multi-layer perception machine to obtain the first The project feature time series association implicit aggregation feature vector corresponding to the project feature time series association implicit representation vector.

[0048] In the embodiment of the present application, specifically, the information transfer aggregation module is used to: use the information transfer aggregation network to process the sequence of the project feature temporal association implicit representation vectors using the following transfer aggregation formula to obtain the project feature temporal aggregation representation vector; wherein the transfer aggregation formula is: ;in, is the first in the sequence of the implicit representation vector of the temporal association of the project features The temporal association implicit representation vector of item features, is the first in the sequence of the implicit representation vector of the temporal association of the project features The temporal association implicit representation vector of item features, is a trainable preset hyperparameter, represents a multi-layer perceptron, is the number of feature vectors in the sequence of implicit representation vectors associated with the project feature temporal sequence, A temporal aggregation representation vector representing the project features.

[0049] In a preferred embodiment, after the sequence of the project feature time series associated implicit representation vectors is passed through the information transmission aggregation network to obtain the project feature time series aggregate representation vector, it further includes: calculating the self-mean matrix and self-standard deviation matrix of the project feature time series aggregate representation vector, wherein the self-mean matrix The value of the position is the first Position eigenvalue and The mean of the position eigenvalues, and the self-standard deviation matrix The value of the position is the first Position eigenvalue and The standard deviation of the positional feature value; matrix multiplying the transposed vector of the project feature time series aggregation representation vector as the row feature vector with the self-mean matrix to obtain the project feature time series aggregation mean mapping vector and matrix multiplying the self-standard deviation matrix with the project feature time series aggregation representation vector to obtain the project feature time series aggregation standard deviation mapping vector; calculating the point sum of the transposed vector of the project feature time series aggregation mean mapping vector and the project feature time series aggregation standard deviation mapping vector to obtain the project feature time series aggregation mean-standard deviation vector; calculating the matrix product of the self-mean matrix and the self-standard deviation matrix, and matrix multiplying the transposed vector of the project feature time series aggregation representation vector with the matrix product to obtain the project feature time series aggregation reference representation vector; and, calculating the point sum of the project feature time series aggregation mean-standard deviation vector and the transposed vector of the project feature time series aggregation reference representation vector to obtain the optimized project feature time series aggregation representation vector.

[0050] In particular, here, taking into account the implicit temporal pattern feature representation of the time series of each group of project features by each project feature temporal association implicit representation vector in the sequence of the project feature temporal association implicit representation vectors, in the case of differences in embedded coding and implicit temporal pattern feature expression of each group of project features, there will also be certain feature distribution differences between the project feature temporal association implicit representation vectors. Therefore, when information transmission and aggregation are performed in the global time domain, there will be local mapping overflow relative to the global time domain aggregation feature space of the project feature temporal aggregation representation vector, thereby affecting the expression effect of the project feature temporal aggregation representation vector.

[0051] Based on this, the self-mean matrix and the self-standard deviation matrix of the group aggregation statistical evaluation of the local distribution of the feature value granularity of each project feature time series aggregation representation vector in the multiple project feature time series aggregation representation vectors are used as the retrieval and response distribution enhancements of each project feature time series aggregation representation vector in the multiple project feature time series aggregation representation vectors, and a reference-free distribution retrieval response framework based on the group aggregation feature distribution open domain of each project feature time series aggregation representation vector in the multiple project feature time series aggregation representation vectors is constructed, so as to avoid the distribution response redundancy caused by the local mapping overflow of each project feature time series aggregation representation vector in the multiple project feature time series aggregation representation vectors through response superposition, so as to improve the mapping effect of each project feature time series aggregation representation vector in the multiple project feature time series aggregation representation vectors in the global time domain aggregation feature space, and enhance the expression effect of each project feature time series aggregation representation vector in the multiple project feature time series aggregation representation vectors, thereby improving the accuracy of the monitoring results of the project multi-dimensional feature time series aggregation representation vectors obtained by processing the multiple project feature time series aggregation representation vectors using a feature aggregator based on a multi-layer perceptron model through a classifier-based project progress intelligent monitor. This approach can provide a more comprehensive status analysis to timely update project progress and provide data-driven decision support for the project management team to develop more effective management strategies.

[0052] In an embodiment of the present application, the feature aggregation module 140 is used to perform feature aggregation on the multiple project feature time series aggregation representation vectors to obtain a project multi-dimensional feature time series aggregation representation vector as a project multi-dimensional feature time series aggregation representation feature. Specifically, in an embodiment of the present application, the feature aggregation module is used to: use a feature aggregator based on a multi-layer perceptron model to process the multiple project feature time series aggregation representation vectors to obtain the project multi-dimensional feature time series aggregation representation vector. It should be understood that, considering that each of the multiple project feature time series aggregation representation vectors reflects the project time series aggregation feature information of a feature dimension. Therefore, in order to aggregate and integrate the project feature time series aggregation representation vectors of multiple feature dimensions to obtain a comprehensive feature representation to characterize the multi-dimensional time series feature information of the project, in the technical solution of the present application, a feature aggregator based on a multi-layer perceptron model is used to process the multiple project feature time series aggregation representation vectors to obtain the project multi-dimensional feature time series aggregation representation vector. Specifically, the multi-layer perceptron model can fuse and map the multiple project feature time series aggregation representation vectors to better capture the complex relationship between the features of each project feature time series aggregation representation, thereby generating a higher-dimensional and richer project multi-dimensional feature time series aggregation representation vector to more comprehensively characterize the overall feature information of the project.

[0053] In an embodiment of the present application, the monitoring result generation module 150 is used to obtain monitoring results based on the time series aggregation representation features of the project multi-dimensional features. Specifically, in an embodiment of the present application, the monitoring result generation module is used to: pass the project multi-dimensional feature time series aggregation representation vector through a classifier-based project progress intelligent monitor to obtain a monitoring result, and the monitoring result is used to indicate whether there is an abnormality in the project progress. That is, the project multi-dimensional feature time series aggregation representation features obtained by transferring and aggregating the multiple project feature time series aggregation representation vectors are used for classification processing, so as to intelligently obtain the monitoring result of whether there is an abnormality in the project progress. Through this method, a more comprehensive status analysis can be provided to timely update the project progress, and data-driven decision support can be provided to the project management team to formulate more effective management strategies.

[0054] In summary, the project status prediction system 100 based on the change of project status characteristics based on the embodiment of the present application is explained, which extracts the time series of the project multidimensional characteristics of the monitored project from the background database, and uses the data processing and analysis technology based on the deep learning neural network to extract and associate the project characteristics of each dimension in the time series of the project multidimensional characteristics, so as to intelligently obtain the monitoring results of whether there are abnormalities in the project progress according to the multidimensional time series aggregation characteristics of each project. Through this method, it is possible to provide a more comprehensive status analysis, update the project progress in a timely manner, and provide data-driven decision support for the project management team to formulate more effective management strategies.

[0055] As described above, the project status prediction system 100 based on changes in project status characteristics according to the embodiment of the present application can be implemented in various wireless terminals, such as a server having a project status prediction algorithm based on changes in project status characteristics. In a possible implementation, the project status prediction system 100 based on changes in project status characteristics according to the embodiment of the present application can be integrated into a wireless terminal as a software module and / or a hardware module. For example, the project status prediction system 100 based on changes in project status characteristics 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 project status prediction system 100 based on changes in project status characteristics can also be one of the many hardware modules of the wireless terminal.

[0056] Alternatively, in another example, the project status prediction system 100 based on changes in project status characteristics and the wireless terminal may also be separate devices, and the project status prediction system 100 based on changes in project status characteristics 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.

[0057] The above specific implementation methods further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above are only specific implementation methods of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A project status prediction system based on changes in project status characteristics, characterized in that: include: A project feature acquisition module is used to extract the time series of project multidimensional features of the monitored project from the background database, wherein the project multidimensional features include time-related features, cost-related features, and progress-related features; A data regularization module, used for performing data regularization on the time series of the multi-dimensional features of the project to obtain time series of multiple groups of project features; A feature time series aggregation module, used to extract multiple project feature time series aggregation representation vectors from the time series of the multiple groups of project features; A feature aggregation module, used for performing feature aggregation on the plurality of project feature time series aggregation representation vectors to obtain a project multi-dimensional feature time series aggregation representation vector as a project multi-dimensional feature time series aggregation representation feature; A monitoring result generation module, used for obtaining monitoring results based on the time series aggregation representation features of the multi-dimensional features of the project; Wherein, the feature time series aggregation module includes: A project feature embedding coding unit, for embedding and coding each project feature in the time series of each group of project features in the time series of the multiple groups of project features to obtain a time series of project feature embedding coding vectors; A time series association implicit pattern feature extraction unit, used for extracting the time series association implicit pattern feature of the project features from the time series of the project feature embedding coding vector to obtain a sequence of project feature time series association implicit representation vectors; An information transfer aggregation unit, used for passing the sequence of the project feature time series associated implicit representation vectors through an information transfer aggregation network to obtain the project feature time series aggregation representation vector; Wherein, the information transmission aggregation unit includes: A time series correlation implicit aggregation feature calculation subunit, used to calculate the project feature time series correlation implicit aggregation feature vector corresponding to each project feature time series correlation implicit representation vector in the sequence of project feature time series correlation implicit representation vectors; an aggregation subunit, configured to sum the project feature temporal association implicit aggregation feature vectors corresponding to the respective project feature temporal association implicit representation vectors in the sequence of the project feature temporal association implicit representation vectors by position to obtain the project feature temporal association aggregation representation vector; The temporal correlation implicit aggregation feature calculation subunit includes: The feature summation secondary subunit is used to calculate the first and last in the sequence of the temporal association implicit representation vector of the item feature. The weighted sum of the implicit representation vectors of the temporal association of the project features is obtained to obtain the implicit sum representation vector of the temporal association of the project features; The feature weighted secondary subunit is used to weight the first feature using the sum of the trainable preset hyperparameter and one as the weighting coefficient. Performing weighted processing on each position feature value in the project feature time series association implicit representation vector to obtain a weighted project feature time series association implicit representation vector; The multi-layer perception secondary subunit is used to process the implicit sum representation vector of the project feature temporal association and the weighted project feature temporal association implicit representation vector by position through a multi-layer perception machine to obtain the first The project feature time series association implicit aggregation feature vector corresponding to the project feature time series association implicit representation vector.

2. The project status prediction system based on project status feature changes according to claim 1, characterized in that: The data regularization module is used to: perform data regularization on the time series of the multi-dimensional features of the project based on the feature dimension to obtain the time series of the multiple groups of project features.

3. The project status prediction system based on project status feature changes according to claim 2, characterized in that: The temporal association implicit pattern feature extraction unit is used to: embed the project features into the time series of the coding vector through a project feature temporal association implicit pattern feature extractor based on a bidirectional LSTM model to obtain a sequence of the project feature temporal association implicit representation vectors.

4. The project status prediction system based on project status feature changes according to claim 3 is characterized in that: The feature aggregation module is used to: use a feature aggregator based on a multi-layer perceptron model to process the multiple project feature time series aggregation representation vectors to obtain the project multi-dimensional feature time series aggregation representation vector.

5. The project status prediction system based on project status feature changes according to claim 4, characterized in that: The monitoring result generation module is used to: pass the project multi-dimensional feature time series aggregation representation vector through a classifier-based project progress intelligent monitor to obtain a monitoring result, and the monitoring result is used to indicate whether there is an abnormality in the project progress.

Images