Innovative project decision prompting method and system based on AI model analysis

Abstract

The invention discloses an innovative project decision prompting method based on AI model analysis, and relates to the technical field of innovative project monitoring, innovative project templates are constructed, a project core label is set for each template, a project execution mapping tree is designed according to project activity types, and each execution node of the mapping tree comprises project execution keywords and monitoring content; analyzing innovation project information through real-time semantic understanding, extracting project core labels and real-time activity progress information, and comparing the project core labels and the real-time activity progress information with templates to call corresponding templates; mapping the real-time progress information to an execution keyword, performing monitoring in combination with corresponding monitoring content, and generating warning content; according to the invention, real-time monitoring and accurate warning of the progress of the innovative project are realized, the intelligent level of project management is effectively improved, the project risk is reduced, and smooth promotion of the innovative project is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of innovation project monitoring, and particularly to a method and system for innovation project decision-making prompting based on AI model analysis. Background Art

[0002] With the rapid development of technology, innovation projects play an increasingly important role in the development of enterprises. The characteristics of innovation projects usually include high complexity, strong uncertainty, involving multi-party collaboration, etc., which pose higher requirements for project management and monitoring. Effective project monitoring can timely discover potential risks, ensure the project progresses according to the plan, and ultimately achieve the expected goals. However, traditional project management methods have many deficiencies when facing innovation projects.

[0003] In the prior art, project monitoring usually relies on manual regular inspections or simple automated systems based on fixed rules. For example, some project management systems evaluate project progress by setting milestones and key performance indicators (KPIs), but this method lacks real-time performance and cannot dynamically adapt to project changes. In addition, when dealing with a large amount of project data, traditional methods often use manual analysis or simple statistical tools, and it is difficult to extract key information from complex data and generate targeted warning content. This results in the lag in discovering potential problems and increases the risk of project failure. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system that can effectively monitor innovation projects.

[0005] The present invention discloses a method for innovation project decision-making prompting based on AI model analysis, including:

[0006] Step S100, constructing an innovation project template, and setting a number of project core tags for each innovation project template. The innovation project template includes project activity types, and each project activity type is set with a project execution mapping tree. Each execution node of the project execution mapping tree is set with a project execution keyword and project monitoring content;

[0007] Step S200, performing real-time semantic understanding on innovation project information, determining the project core tags and real-time activity progress information of the innovation project, comparing the project core tags in the innovation project template, and determining the called innovation project template;

[0008] Step S300, based on the real-time activity progress information, determining the project execution keywords mapped in the project execution mapping tree, and monitoring the real-time activity progress information based on the project monitoring content corresponding to the project execution keywords, and generating warning content.

[0009] In some embodiments disclosed by the present invention, the method for determining the project core tags of an innovation project template includes:

[0010] Step S101: Obtain historical innovation project data, perform text preprocessing on each piece of innovation project data, and use a keyword recognition algorithm to recognize keywords in the innovation project data to determine a number of candidate keywords;

[0011] Step S102: Compare and analyze each candidate keyword with a number of preset project keyword sets to determine the project keyword set corresponding to each candidate keyword, denoted as the concerned keyword set;

[0012] Step S103: Judge the number of candidate keywords mapped in each concerned keyword set, and sort the concerned keyword sets based on the number of candidate keywords mapped to obtain a sequence of concerned keyword sets;

[0013] Step S104: Configure an initial core degree for the candidate keywords based on the order of the earliest concerned keyword set to which the candidate keywords belong, and the proportion of candidate keywords in the concerned keyword set to the total number of candidate keywords. Correct the initial core degree of the candidate keywords based on the average core degree of other candidate keywords in the same concerned keyword set to obtain a reference core degree;

[0014] Step S105: Retain the candidate keywords based on their reference core degrees and recognize them as project core tags.

[0015] In some embodiments disclosed by the present invention, the method for configuring an initial core degree for candidate keywords includes:

[0016] Step S1031: For different orders of the sequence of concerned keyword sets, a basic core degree H0 is set. Configure the basic core degree for the candidate keywords based on the order of the earliest concerned keyword set to which the candidate keywords belong;

[0017] Step S1032: A keyword proportion array [j1, j2, j3,..., jn] is set for the candidate keyword proportion, where 0 < j1 < j2 < j3 <... < jn, and an initial adjustment coefficient array [R1, R2, R3,..., Rn] is set for the keyword proportion array, where 0 < R1 < R2 < R3 <... < Rn. Judge the keyword proportion interval to which the candidate keyword proportion j0 belongs:

[0018] If j0 ≤ j1, determine that the initial adjustment coefficient of the basic core degree is R1, then the initial core degree H1 of the candidate keyword = R1 × H0;

[0019] If j1 < j0 ≤ j2, then determine that the initial adjustment coefficient of the basic core degree is R2, and the initial core degree H1 of the alternative keyword is H1 = R2 × H0;

[0020] If j2 < j0 ≤ j3, then determine that the initial adjustment coefficient of the basic core degree is R3, and the initial core degree H1 of the alternative keyword is H1 = R3 × H0; ...;

[0022] If jn - 1 < j0 ≤ jn, then determine that the initial adjustment coefficient of the basic core degree is Rn, and the initial core degree H1 of the alternative keyword is H1 = Rn × H0.

[0023] In some embodiments disclosed by the present invention, the method for correcting the initial core degree of the alternative keyword includes:

[0024] Step S1041, a core degree array [h1, h2, h3,..., hn] is set for the average core degree of the alternative keyword, where h1 < h2 < h3 <... < hn, and a secondary adjustment coefficient array [F1, F2, F3,..., Fn] is set for the core degree array, where F1 < F2 < F3 <... < Fn. Determine the core degree interval to which the average core degree H3 belongs:

[0025] If H3 ≤ h1, then determine that the initial adjustment coefficient of the basic core degree is F1, and the reference core degree H2 of the alternative keyword is H2 = F1 × H1;

[0026] If h1 < H3 ≤ h2, then determine that the initial adjustment coefficient of the basic core degree is F2, and the reference core degree H2 of the alternative keyword is H2 = F2 × H1;

[0027] If h2 < H3 ≤ h3, then determine that the initial adjustment coefficient of the basic core degree is F3, and the reference core degree H2 of the alternative keyword is H2 = F3 × H1; ...;

[0029] If hn - 1 < H3 ≤ hn, then determine that the initial adjustment coefficient of the basic core degree is Fn, and the reference core degree H2 of the alternative keyword is H2 = Fn × H1.

[0030] In some embodiments disclosed by the present invention, and based on the project monitoring content corresponding to the project execution keyword, the method for monitoring the real-time activity progress information includes:

[0031] Step S301, extract the real-time execution keywords from the real-time activity progress information, determine the time node corresponding to each real-time execution keyword, and sort the real-time execution keywords based on the time series relationship to obtain a real-time execution keyword sequence;

[0032] Step S302: Compare the real-time execution keyword sequence with the project execution mapping tree. If the real-time execution keywords in the real-time execution keyword sequence are compared with the execution nodes in the project execution mapping tree in sequence, and several continuously mapped execution nodes are determined, the continuous execution nodes are recorded as the mapped execution branch;

[0033] Step S303: Sort the mapped execution branches based on the length of the mapped execution branch, and select the project monitoring content corresponding to the top several mapped execution branches to monitor the real-time activity progress information.

[0034] In some embodiments disclosed by the present invention, the method for monitoring the real-time activity progress information includes:

[0035] Step S3031: Determine the progress factor in the real-time activity progress information based on the project monitoring content, and judge the change parameter information within a preset range near the information segment corresponding to the progress factor, and judge the change parameter information corresponding to the progress factor;

[0036] Step S3032: Construct a progress parameter change model corresponding to the progress factor based on the change parameter information corresponding to the progress factor. The progress parameter change model is used to express the change of the progress parameter at different time nodes, calculate the integral of the parameter change of the progress parameter change model over a continuous time period to obtain a continuous parameter change integral. If the continuous parameter change integral does not meet the preset standard, alarm the corresponding progress parameter change model;

[0037] Step S3033: Record the combination of the alarmed progress parameter change models as the key monitoring change model group, and analyze the key monitoring change model group to generate warning content.

[0038] In some embodiments disclosed by the present invention, the method for analyzing the key monitoring change model group to generate warning content includes:

[0039] Step S3034: Divide the progress parameter change model successively according to the equivalent parameter change interval, and determine the time length of each divided model section, which is recorded as the time length for model comparison. Record the combination of the time lengths for model comparison belonging to the same progress parameter change model as the time length group for model comparison;

[0040] Step S3035: Based on the equivalence of the progress parameter change models in the key monitoring change model group and the equivalence of the time length group for model comparison corresponding to each progress parameter change model, screen out the corresponding historical monitoring change model group from several preset historical monitoring change model groups, and recognize the warning content corresponding to the historical monitoring change model group as the warning content of the key monitoring change model group.

[0041] In some embodiments disclosed by the present invention, there is also disclosed an innovation project decision-making prompt system based on AI model analysis, including:

[0042] A first module, configured to construct an innovation project template, and set a number of project core tags for each innovation project template. The innovation project template includes project activity types, and each project activity type is set with a project execution mapping tree. Each execution node of the project execution mapping tree is set with a project execution keyword and project monitoring content;

[0043] A second module, configured to perform real-time semantic understanding on innovation project information, determine the project core tags and real-time activity progress information of the innovation project, compare the project core tags in the innovation project template, and determine the called innovation project template;

[0044] A third module, configured to determine the project execution keywords mapped in the project execution mapping tree based on the real-time activity progress information, and monitor the real-time activity progress information based on the project monitoring content corresponding to the project execution keywords, and generate warning content.

[0045] The present invention discloses an innovation project decision-making prompt method based on AI model analysis, which relates to the technical field of innovation project monitoring. An innovation project template is constructed, project core tags are set for each template, and a project execution mapping tree is designed according to the project activity type. Each execution node of the mapping tree includes a project execution keyword and monitoring content; the innovation project information is analyzed through real-time semantic understanding to extract the project core tags and real-time activity progress information, and compared with the template to call the corresponding template; based on the real-time progress information, it is mapped to the execution keyword, and the corresponding monitoring content is combined for monitoring and warning content is generated; the present invention realizes the real-time monitoring and accurate warning of the progress of innovation projects, effectively improves the intelligent level of project management, reduces project risks, and ensures the smooth progress of innovation projects.

[0046] The technical solution of the present invention will be further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a method step diagram of the innovation project decision-making prompt method based on AI model analysis disclosed in the embodiments of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0048] The technical solution of the present invention will be further described below through the accompanying drawings and embodiments.

[0049] The technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and should not be construed as limiting the protection scope of the present invention. Skilled technicians in this field can make some non-essential improvements and adjustments based on the content of the present invention below. In the present invention, unless otherwise clearly specified and limited, the technical terms used in the present invention should have the ordinary meaning understood by those skilled in the art of the present invention.

[0050] Embodiment:

[0051] The present invention discloses an innovative project decision-making prompt method based on AI model analysis. Refer to Figure 1 , including:

[0052] Step S100, constructing an innovative project template, and setting several project core tags for each innovative project template. The innovative project template includes project activity types, and each project activity type is set with a project execution mapping tree. Each execution node of the project execution mapping tree is set with a project execution keyword and project monitoring content.

[0053] The principle of this step is to establish a standardized monitoring framework for innovative projects through structured design, that is, the "innovative project template", aiming to decompose the complexity of the project into manageable units and provide a basis for subsequent dynamic monitoring. The system first constructs multiple templates according to historical project data or expert experience. Each template is for a specific type of innovative project. For example, a technology R & D project may include "technology R & D" as the core tag. The "project activity type" is further defined in the template, such as "requirements analysis", "prototype development", "test verification", etc. Each activity type corresponds to a "project execution mapping tree". The mapping tree is a tree structure that reflects the execution process in the activity. Each node represents a task stage. For example, "complete the requirements document" is a node, and this node is bound with a "project execution keyword" (such as "requirements confirmation") and "project monitoring content" (such as "the completion time does not exceed 5 days"). For example, assume a template for a software development project, and its mapping tree may include a node chain from "requirements analysis" to "code writing" and then to "testing". Each node has a keyword (such as "code submission") and monitoring content (such as "the code quality needs to reach 90%"). In this way, the template not only sorts out the execution logic of the project, but also presets a standardized evaluation basis for real-time monitoring, ensuring that the subsequent steps can be quickly called and applied.

[0054] Step S200, performing real-time semantic understanding on the innovative project information, determining the project core tags and real-time activity progress information of the innovative project, and comparing the project core tags in the innovative project template to determine the called innovative project template.

[0055] The principle of this step is to use natural language processing (NLP) and semantic analysis in AI technology to intelligently analyze real-time project information, match the characteristics of the current project with pre-built templates, and thus determine the applicable monitoring framework. After the system receives project information (such as a meeting record "Prototype design completed this week" or a log "Requirements confirmed"), it performs semantic understanding through an NLP model (such as BERT), extracts the "project core label" (such as "Prototype development") and the "real-time activity progress information" (such as "50% of the prototype design has been completed"). Subsequently, the system compares the extracted core label with the labels in the template library. For example, if the input information mentions "Technology R & D" and "Prototype", it may match a template with "Technology R & D" as the core label. The comparison process may be based on similarity calculation, such as calculating the cosine similarity between the input label and the template label. If the similarity exceeds a threshold (such as 0.9), the template is called. For example, assume that a project team submits a progress report of "AI algorithm optimization in progress". After system analysis, "AI optimization" is extracted as the core label and matches a template with the theme of "Algorithm R & D", thus providing an execution mapping tree for subsequent monitoring. This mechanism ensures that real-time information can be quickly associated with the appropriate template to achieve dynamic adaptation.

[0056] Step S300: Based on the real-time activity progress information, determine the project execution keywords mapped in the project execution mapping tree, and monitor the real-time activity progress information based on the project monitoring content corresponding to the project execution keywords, and generate warning content.

[0057] The principle of this step is to accurately map the real-time project progress to the execution mapping tree in the template, and dynamically analyze the progress based on the preset monitoring content, and finally generate a risk warning. The system first extracts the "real-time execution keywords" from the real-time activity progress information. For example, it extracts "testing completed" from "50% of the testing phase has been completed", and then matches it with the node keywords in the mapping tree (such as "testing passed") to locate the corresponding execution node. After positioning, the system calls the "project monitoring content" of this node for evaluation. For example, the monitoring content may be "the testing completion rate should reach 100% and the time should not exceed 10 days". If only 50% has been completed and 8 days have passed, an anomaly detection is triggered. The generation of the warning content is based on preset rules or AI analysis. For example, it calculates the degree of progress deviation and outputs a warning such as "the testing progress is lagging behind, and it is expected to be delayed by 2 days". For example, assume that the progress information of a project shows that "the code writing has been submitted, but the quality is only 85%". The system maps to the "code submission" node and finds that its monitoring content requires "the quality should reach 90%", so it generates a warning "the code quality does not meet the standard, it is recommended to optimize". In this way, the system realizes the complete process from keyword matching of real-time data to monitoring evaluation and then to warning generation, ensuring that project managers can timely grasp the risks and take actions.

[0058] In some embodiments disclosed by the present invention, the method for determining the project core label of the innovative project template includes:

[0059] Step S101, obtaining historical innovative project data, performing text preprocessing on each piece of innovative project data, and using a keyword recognition algorithm to recognize keywords in the innovative project data to determine a number of alternative keywords.

[0060] The principle of this step is to extract key information from historical innovation project data through text mining technology, laying a data foundation for the determination of subsequent core tags. The system first collects a large amount of historical project data, such as project plans, progress reports, or summary documents, which reflect the characteristics and implementation processes of innovation projects. Then, text preprocessing is performed on the data, including word segmentation, removal of stop words (such as "of", "is"), stemming, etc., to standardize the text format. Next, keyword recognition algorithms (such as TF-IDF, TextRank, or LDA) are used to analyze the text to identify high-frequency or high-importance words as candidate keywords. These keywords represent the concepts or themes that repeatedly appear in the project. For example, suppose a historical report of a drone R & D project is analyzed, and candidate keywords such as "flight test", "battery optimization", and "cost control" are extracted after preprocessing. These keywords may be the core concerns during project implementation. The system quantifies their occurrence frequency or semantic importance through algorithms to ensure that the extraction results are representative. This process provides data-driven materials for subsequent tag screening, echoing the goal of constructing an innovation project template throughout the text and providing a candidate basis for the core tags of the template.

[0061] Step S102: Compare and analyze each candidate keyword with several preset project keyword sets to determine the project keyword set corresponding to each candidate keyword, denoted as the focus keyword set.

[0062] The principle of this step is to match the candidate keywords with the preset project keyword sets, classify the theme areas to which each keyword belongs, and thus form the focus keyword set, providing a structured basis for subsequent importance assessment. The system pre-defines several project keyword sets, such as "technology category" (including "algorithm", "test", etc.), "management category" (including "resource", "progress", etc.), which are constructed based on industry experience or expert knowledge. Then, the candidate keywords extracted in step S101 are compared with these preset sets one by one, and their semantic relevance is analyzed (possibly through word vector similarity or rule matching) to determine the keyword set corresponding to each candidate keyword, denoted as the focus keyword set. For example, in the drone project, "flight test" and "battery optimization" may be classified into the "technology category", and "cost control" into the "management category". This process is similar to a classification task, aiming to organize scattered keywords into meaningful groups to ensure that subsequent analysis can focus on specific theme areas. This is consistent with the logic of dividing the "project activity type" when constructing the template in step S100 throughout the text and provides support for the structured design of the template.

[0063] Step S103: Judge the number of candidate keywords mapped to each focus keyword set, and sort the focus keyword sets based on the number of candidate keywords mapped to them to obtain the focus keyword set sequence.

[0064] The principle of this step is to quantify and rank the importance of the keyword sets of interest by statistical analysis of their coverage, providing a basis for prioritization in subsequent core level configuration. The system counts the number of alternative keywords included in each keyword set of interest. For example, the "technology category" contains 8 keywords (such as "flight test", "battery optimization", etc.), and the "management category" contains 3 (such as "cost control", etc.). Then, they are sorted from most to least in terms of quantity to generate a sequence of keyword sets of interest. The sorting reflects the universality and representativeness of the keyword sets in historical data. Sets with a larger quantity (such as the "technology category") usually correspond to the main features of the project and may be closer to the requirements of the core tags. For example, in a drone project, the "technology category" ranks first in the sequence because it contains more keywords, followed by the "management category". This serialization process provides an evaluation dimension for the subsequent step S104, ensuring that the screening of core tags depends not only on individual keywords but also on the overall importance of their affiliated fields. This idea of quantitative sorting is similar to the logic of sorting based on the length of the mapping branch in step S300 throughout the text, reflecting the system's emphasis on data prioritization.

[0065] Step S104: Configure the initial core level for the alternative keywords based on the order of the keyword set of interest that the alternative keyword belongs to and the proportion of the alternative keyword in the keyword set of interest to the total number of alternative keywords. Then, correct the initial core level of the alternative keyword based on the average core level of other alternative keywords in the same keyword set of interest to obtain the reference core level.

[0066] The principle of this step is to quantitatively evaluate the importance of alternative keywords through multiple dimensions, configure the initial core level by combining the sorting and proportion of the affiliated keyword set, and correct it through the average value of the same group of keywords, finally obtaining a more accurate reference core level. The system first calculates the initial core level based on the order of the keyword set of interest that the alternative keyword belongs to (such as the "technology category" ranking first) and the proportion of the alternative keyword in this set to the total keywords (such as "flight test" accounting for 25% of the "technology category").

[0067] Step S105: Retain the alternative keywords based on their reference core level and identify them as the core tags of the project.

[0068] In some embodiments disclosed in the present invention, the method for configuring the initial core level for alternative keywords includes:

[0069] Step S1031: For different orders of the keyword set sequence of interest, set a basic core level H0, and configure the basic core level for the alternative keywords based on the order of the keyword set of interest that the alternative keyword belongs to.

[0070] Step S1032: Set a keyword ratio array [j1, j2, j3,..., jn] for the alternative keyword ratios, where 0 < j1 < j2 < j3 <... < jn. Set an initial adjustment coefficient array [R1, R2, R3,..., Rn] for the keyword ratio array, where 0 < R1 < R2 < R3 <... < Rn. Determine the keyword ratio interval to which the alternative keyword ratio j0 belongs:

[0071] If j0 ≤ j1, determine that the initial adjustment coefficient for the basic core degree is R1, then the initial core degree H1 of the alternative keyword is H1 = R1 × H0.

[0072] If j1 < j0 ≤ j2, determine that the initial adjustment coefficient for the basic core degree is R2, then the initial core degree H1 of the alternative keyword is H1 = R2 × H0.

[0073] If j2 < j0 ≤ j3, determine that the initial adjustment coefficient for the basic core degree is R3, then the initial core degree H1 of the alternative keyword is H1 = R3 × H0.

[0074] ...

[0075] If jn - 1 < j0 ≤ jn, determine that the initial adjustment coefficient for the basic core degree is Rn, then the initial core degree H1 of the alternative keyword is H1 = Rn × H0.

[0076] In some embodiments disclosed in the present invention, the method for correcting the initial core degree of the alternative keyword includes:

[0077] Step S1041: Set a core degree array [h1, h2, h3,..., hn] for the average core degree of the alternative keyword, where h1 < h2 < h3 <... < hn. Set a secondary adjustment coefficient array [F1, F2, F3,..., Fn] for the core degree array, where F1 < F2 < F3 <... < Fn. Determine the core degree interval to which the average core degree H3 belongs:

[0078] If H3 ≤ h1, determine that the initial adjustment coefficient for the basic core degree is F1, then the reference core degree H2 of the alternative keyword is H2 = F1 × H1.

[0079] If h1 < H3 ≤ h2, determine that the initial adjustment coefficient for the basic core degree is F2, then the reference core degree H2 of the alternative keyword is H2 = F2 × H1.

[0080] If h2 < H3 ≤ h3, determine that the initial adjustment coefficient for the basic core degree is F3, then the reference core degree H2 of the alternative keyword is H2 = F3 × H1. ...;

[0082] If hn-1 < H3 ≤ hn, the initial adjustment coefficient for determining the basic core degree is Fn, and the reference core degree H2 of the alternative keyword is H2 = Fn × H1.

[0083] In some embodiments disclosed by the present invention, and based on the project monitoring content corresponding to the project execution keyword, the method for monitoring the real-time activity progress information includes:

[0084] Step S301, extract real-time execution keywords from the real-time activity progress information, determine the time node corresponding to each real-time execution keyword, and sort the real-time execution keywords based on the time series relationship to obtain a real-time execution keyword sequence.

[0085] The principle of this step is to extract key descriptions from the project progress information through real-time text analysis and organize them into an ordered sequence in combination with the time dimension, providing structured input for subsequent mapping and monitoring. The system first receives real-time activity progress information (such as the log "Requirement analysis completed on March 10th, code submitted on March 12th"), and uses natural language processing techniques (such as keyword extraction algorithms or named entity recognition) to extract "real-time execution keywords" from it, such as "requirement analysis" and "code submission". Then, associate each keyword with its corresponding time node (e.g., "requirement analysis" corresponds to March 10th, "code submission" corresponds to March 12th), and sort them based on the time series relationship to form a "real-time execution keyword sequence" (such as "requirement analysis → code submission"). For example, assume that the progress report of a software development project shows "Prototype design completed on March 11th, testing started on March 13th", and the system extracts two keywords "prototype design" and "testing", and sorts them by time as "prototype design → testing". This process converts scattered real-time information into an ordered keyword sequence, ensuring that subsequent steps can match the template in the execution order, while retaining the time dimension for monitoring timeliness, which is consistent with the goal of dynamic monitoring in the overall solution.

[0086] Step S302, compare the real-time execution keyword sequence with the project execution mapping tree. If the real-time execution keywords in the real-time execution keyword sequence are compared with the execution nodes in the project execution mapping tree in sequence and several continuously mapped execution nodes are determined, the continuous execution nodes are recorded as the mapped execution branch.

[0087] The principle of this step is to match the real-time execution keyword sequence with the preset project execution mapping tree, identify the execution path corresponding to the current progress, and lock specific nodes for subsequent monitoring. The project execution mapping tree is a tree structure that contains multiple execution nodes, and each node is bound with keywords (such as "requirement confirmation", "code submission") and monitoring content. The system compares the real-time execution keyword sequence (such as "requirement analysis → code submission") generated in step S301 with the mapping tree one by one to check whether the keywords in the sequence match the nodes of the tree in order. If the match is successful and the nodes are continuous, it is recorded as the "mapped execution branch". For example, in the mapping tree of a software development template, the node chain may be "requirement confirmation → code writing → code submission", and the real-time sequence "requirement analysis → code submission" matches the branch "requirement confirmation → code submission" (assuming that "requirement analysis" is semantically similar to "requirement confirmation"). If the sequence is "prototype design → testing", it may match the branch "prototype development → test verification". The continuity requirement ensures that the branch reflects the actual execution process and avoids scattered matching. This mechanism echoes the design logic of the mapping tree in step S100 and provides accurate positioning for monitoring.

[0088] Step S303: Based on the length of the mapped execution branch, sort the mapped execution branches, and select the project monitoring content corresponding to the top several mapped execution branches respectively to monitor the real-time activity progress information.

[0089] The principle of this step is to dynamically analyze and detect risks for the real-time progress by evaluating the coverage of the mapped execution branch, preferentially selecting the critical path and invoking its monitoring content. The system first calculates the length of each mapped execution branch, that is, the number of continuously matched nodes. For example, the length of "requirement confirmation → code submission" is 2, and the length of "prototype development → test verification → release" is 3. Then, sort them in descending order of length. The longer the branch, the more complete the execution path usually represents, reflecting the core part of the project progress. Next, select the top several branches (such as the top 2), and extract their corresponding "project monitoring content" for monitoring. For example, assume that the monitoring content of the branch "code writing → code submission" is "The code submission should be completed within 5 days with a quality of 90%". The real-time information shows that "submitted on March 12th with a quality of 85%". The system detects that the quality does not meet the standard. The monitoring result may trigger a warning, such as "The code quality is lower than expected". For example, if a project has branches "prototype design → testing" (length 2) and "requirement analysis" (length 1), the system preferentially monitors the former and finds that "The testing progress is only 50% and it is expected to be delayed by 3 days", thus generating a corresponding warning. This process focuses on the critical progress through length sorting, which is consistent with the logic of generating warning content in step S300 of the overall solution, ensuring the pertinence and effectiveness of monitoring.

[0090] In some embodiments disclosed by the present invention, a method for monitoring real-time activity progress information includes:

[0091] Step S3031: Based on the project monitoring content, determine the progress factors in the real-time activity progress information, and determine the change parameter information within a preset range near the information segment corresponding to the progress factor, and determine the change parameter information corresponding to the progress factor.

[0092] The principle of this step is to extract key progress factors and quantify their change characteristics by analyzing the real-time activity progress information and the preset project monitoring content, providing a data basis for subsequent modeling and monitoring. First, based on the "project monitoring content" of the nodes in the project execution mapping tree (such as "completion time not exceeding 5 days", "quality needs to reach 90%"), the system identifies the "progress factors" from the real-time progress information, that is, the key variables affecting project execution, such as "progress", "resource utilization rate", "quality". Then, the system analyzes the data within a preset range near the information segment corresponding to the progress factor (such as "50% of the test completed on March 12"), and extracts the "change parameter information", such as the progress change rate, resource consumption trend, etc. For example, assuming the monitoring content is "the test phase needs to be completed in 10 days", and the real-time information shows "30% completed on March 10 and 50% completed on March 12", the system determines "progress" as the progress factor and calculates the change parameter as "daily progress increment" (the increment from March 10 to 12 is 20%, with an average of 10% per day). This process transforms the abstract monitoring target into measurable parameters through quantitative analysis, ensuring that the subsequent steps can be evaluated based on specific data, which is consistent with the accuracy requirements of dynamic monitoring in the overall solution.

[0093] Step S3032: Based on the change parameter information corresponding to the progress factor, construct a progress parameter change model corresponding to the progress factor. The progress parameter change model is used to express the change of the progress parameter at different time nodes, calculate the integral of the parameter change of the progress parameter change model over a continuous time period to obtain the continuous parameter change integral. If the continuous parameter change integral does not meet the preset standard, an alarm is issued for the corresponding progress parameter change model.

[0094] Step S3033: Denote the combination of the alarmed progress parameter change models as the key monitoring change model group, and analyze the key monitoring change model group to generate warning content.

[0095] The principle of this step is to integrate the change models that trigger alarms into key monitoring objects, generate specific warning content by analyzing their combined characteristics, and provide decision-making support for project management. The system combines the progress parameter change models of the alarms in step S3032 (such as the "test progress lag model") into a "key monitoring change model group". For example, if there are both a "resource consumption exceeding the standard model" and a "progress lag model" at the same time, a model group is formed. Then, the system analyzes the characteristics of this group, such as the alarm frequency, the influence range, or the mutual relevance, and generates warning content. For example, if it is found that "test progress lag" is related to "insufficient resources", the warning content may be "Insufficient resources in the test phase lead to schedule delays. It is recommended to increase the manpower". For example, assume that a project alarms in both the "code writing" and "test" phases, and the model group shows "low code quality" and "slow test progress". The system generates a warning "Unqualified code quality affects the test progress. It is recommended to optimize the code and extend the test time". This process integrates multi-dimensional anomalies and refines key information, which is consistent with the intelligent goal of generating warning content in the overall solution, ensuring that managers can quickly understand and respond to risks.

[0096] In some embodiments disclosed by the present invention, the method for analyzing the key monitoring change model group and generating warning content includes:

[0097] Step S3034, divide the progress parameter change models successively according to the equivalent parameter change interval, and determine the time length of each segmented model section, which is denoted as the time length for model comparison. The combination of the time lengths for model comparison belonging to the same progress parameter change model is denoted as the time length group for model comparison.

[0098] The principle of this step is to perform a standardized segmentation of the progress parameter change model in the time dimension, quantify the distribution law of its change characteristics, and provide a consistent basis for subsequent comparison with historical data. First, for each progress parameter change model in the key monitoring change model group (such as the "test progress lag model"), the system divides it according to the "equivalent parameter change interval", that is, with a fixed change amount (such as a 10% increase in progress each time) as the unit, and divides the time axis of the model into several sections. For example, assuming a model shows "progress 30% on March 10, 50% on March 12, and 70% on March 15", it is divided into four sections of "30%-40%", "40%-50%", "50%-60%", and "60%-70%" at 10% intervals, and the time length of each section is calculated respectively (such as "30%-40%" takes 1 day, "40%-50%" takes 1 day, "50%-60%" takes 2 days, and "60%-70%" takes 1 day), which is recorded as the "time length for model comparison" (1 day, 1 day, 2 days, 1 day). Then, all the time lengths of the same model are combined into a "time length group for comparison" (such as [1, 1, 2, 1]). This process simplifies the complex dynamic changes into discrete time feature sequences through standardized segmentation, ensuring comparability between different models. For example, if another model "resource consumption" is divided at 10% consumption, the result is [2, 1, 3], and the system can quantify its change rhythm through the time length group, echoing the logic based on historical data analysis in the overall plan.

[0099] Step S3035: Based on the equivalence of the progress parameter change models in the key monitoring change model group and the equivalence of the corresponding time length groups for comparison of each progress parameter change model, screen out the corresponding historical monitoring change model group from several preset historical monitoring change model groups, and identify the warning content corresponding to the historical monitoring change model group as the warning content of the key monitoring change model group.

[0100] The principle of this step is to generate applicable warning content by comparing the change characteristics and time distribution similarities between the current model group and the historical model group, using historical experience to improve the efficiency and accuracy of analysis. The system first analyzes the "equivalence of progress parameter change models" of each model in the key monitored change model group (i.e., whether the change trends are similar, such as both being linear growth) and the "equivalence of comparison time length groups" (i.e., whether the time length sequences are close, such as [1,1,2,1] being similar to [1,1,2,2]). Then, it filters out matching items in the preset historical monitored change model group database. For example, assume the current model group includes a "test progress model" (time length group [1,1,2,1]) and a "resource consumption model" ([2,1,3]), and there is a model group record in the historical database as "test progress" ([1,1,2,2]) and "resource consumption" ([2,1,2]). Since their change trends and time length groups are highly similar, this historical model group is selected. Next, the system directly determines the warning content corresponding to the historical model group (such as "Test delay due to insufficient resources, it is recommended to increase manpower") as the warning content for the current model group. For example, if the historical warning is "The progress lags behind by 2 days, the plan needs to be adjusted" and the current situation is similar, the system outputs the same suggestion. This mechanism reuses historical experience through pattern matching, which is consistent with the goal of intelligently generating warnings in the overall solution, avoiding repeated analysis and improving the response speed.

[0101] In some embodiments disclosed by the present invention, there is also disclosed an innovative project decision-making prompt system based on AI model analysis, including:

[0102] A first module for constructing an innovative project template, and setting a number of project core tags for each innovative project template. The innovative project template includes project activity types, and each project activity type is set with a project execution mapping tree. Each execution node of the project execution mapping tree is set with a project execution keyword and project monitoring content;

[0103] A second module for performing real-time semantic understanding on the innovative project information, determining the project core tags and real-time activity progress information of the innovative project, comparing the project core tags in the innovative project template, and determining the called innovative project template;

[0104] A third module for determining the project execution keywords mapped in the project execution mapping tree based on the real-time activity progress information, and monitoring the real-time activity progress information based on the project monitoring content corresponding to the project execution keywords to generate warning content.

[0105] The present invention discloses an innovative project decision-making prompt method based on AI model analysis, which relates to the technical field of innovative project monitoring. An innovative project template is constructed, a project core tag is set for each template, and a project execution mapping tree is designed according to the project activity type. Each execution node of the mapping tree contains project execution keywords and monitoring content. The innovative project information is analyzed through real-time semantic understanding to extract the project core tag and real-time activity progress information, and compared with the template to call the corresponding template. Based on the real-time progress information, it is mapped to the execution keywords, and the corresponding monitoring content is combined for monitoring and warning content is generated. The present invention realizes the real-time monitoring and accurate warning of the progress of innovative projects, effectively improves the intelligent level of project management, reduces project risks, and ensures the smooth progress of innovative projects.

[0106] Through the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by hardware or by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various implementation scenarios of the present invention.

[0107] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that they can still modify or equivalently replace the technical solutions of the present invention, and these modifications or equivalent replacements cannot make the modified technical solutions deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. An innovative project decision-making prompt method based on AI model analysis, characterized in that, Including: Step S100: Construct an innovation project template, and set a number of project core tags for each innovation project template. The innovation project template includes project activity types, and each project activity type is set with a project execution mapping tree. Each execution node of the project execution mapping tree is set with a project execution keyword and project monitoring content; Step S200: Conduct real-time semantic understanding on innovation project information, determine the project core tags and real-time activity progress information of the innovation project, compare the project core tags in the innovation project template, and determine the called innovation project template; Step S300: Based on the real-time activity progress information, determine the project execution keywords mapped in the project execution mapping tree, and monitor the real-time activity progress information based on the project monitoring content corresponding to the project execution keywords to generate warning content.

2. The innovative project decision-making prompt method based on AI model analysis according to claim 1, wherein The method for determining the project core tags of the innovation project template includes: Step S101: Obtain historical innovation project data, conduct text preprocessing on each innovation project data, and use a keyword recognition algorithm to recognize keywords in the innovation project data to determine a number of alternative keywords; Step S102: Compare and analyze each alternative keyword with a number of preset project keyword sets to determine the project keyword set corresponding to each alternative keyword, denoted as the focus keyword set; Step S103: Judge the number of alternative keywords mapped in each focus keyword set, and sort the focus keyword sets based on the number of alternative keywords mapped to obtain a focus keyword set sequence; Step S104: Based on the order of the earliest focus keyword set to which the alternative keyword belongs, and the proportion of the alternative keyword in the focus keyword set to the total number of alternative keywords, configure an initial core degree for the alternative keyword. Based on the average core degree of other alternative keywords in the same focus keyword set, correct the initial core degree of the alternative keyword to obtain a reference core degree; Step S105: Retain the alternative keywords based on their reference core degrees and recognize them as project core tags.

3. The innovative project decision-making prompt method based on AI model analysis according to claim 2, wherein, The method for configuring the initial core degree for alternative keywords includes: Step S1031: For different orders of the focus keyword set sequence, set a basic core degree H0, and configure the basic core degree for the alternative keyword based on the order of the earliest focus keyword set to which the alternative keyword belongs; Step S1032: Set a keyword ratio array [j1, j2, j3,..., jn] for the alternative keyword ratio, where 0 < j1 < j2 < j3 <... < jn, and set an initial adjustment coefficient array [R1, R2, R3,..., Rn] for the keyword ratio array, where 0 < R1 < R2 < R3 <... < Rn. Judge the keyword ratio interval to which the alternative keyword ratio j0 belongs: If j0 ≤ j1, determine that the initial adjustment coefficient of the basic core degree is R1, then the initial core degree H1 of the alternative keyword = R1 × H0; If j1 < j0 ≤ j2, then determine that the initial adjustment coefficient of the basic core degree is R2, and the initial core degree H1 of the alternative keyword = R2 × H0; If j2 < j0 ≤ j3, then determine that the initial adjustment coefficient of the basic core degree is R3, and the initial core degree H1 of the alternative keyword = R3 × H0; ...； If jn-1 < j0 ≤ jn, then determine that the initial adjustment coefficient of the basic core degree is Rn, and the initial core degree H1 of the alternative keyword = Rn × H0.

4. The innovative project decision-making prompt method based on AI model analysis according to claim 2, characterized in that The methods for correcting the initial core degree of the alternative keyword include: Step S1041, for the average core degree of the alternative keyword, there is a core degree array [h1, h2, h3,..., hn] set, where h1 < h2 < h3 <... < hn, and for the core degree array, a secondary adjustment coefficient array [F1, F2, F3,..., Fn] is set, where F1 < F2 < F3 <... < Fn, and determine the core degree interval to which the average core degree H3 belongs: If H3 ≤ h1, then determine that the initial adjustment coefficient of the basic core degree is F1, and the reference core degree H2 of the alternative keyword = F1 × H1; If h1 < H3 ≤ h2, then determine that the initial adjustment coefficient of the basic core degree is F2, and the reference core degree H2 of the alternative keyword = F2 × H1; If h2 < H3 ≤ h3, then determine that the initial adjustment coefficient of the basic core degree is F3, and the reference core degree H2 of the alternative keyword = F3 × H1; ...； If hn-1 < H3 ≤ hn, then determine that the initial adjustment coefficient of the basic core degree is Fn, and the reference core degree H2 of the alternative keyword = Fn × H1.

5. The innovative project decision-making prompt method based on AI model analysis according to claim 1, characterized in that And the method for monitoring the real-time activity progress information based on the project monitoring content corresponding to the project execution keyword includes: Step S301, extract the real-time execution keywords from the real-time activity progress information, and determine the time node corresponding to each real-time execution keyword. Based on the time series relationship, sort the real-time execution keywords to obtain the real-time execution keyword sequence; Step S302, compare the real-time execution keyword sequence with the project execution mapping tree. If the real-time execution keywords in the real-time execution keyword sequence are compared with the execution nodes in the project execution mapping tree in sequence, and several continuously mapped execution nodes are determined, and the continuous execution nodes are recorded as the mapped execution branches; Step S303, sort the mapped execution branches based on the length of the mapped execution branches, and select the project monitoring content corresponding to the first several mapped execution branches to monitor the real-time activity progress information.

6. The innovative project decision-making prompt method based on AI model analysis according to claim 5, characterized in that, The method for monitoring the real-time activity progress information includes: Step S3031, determine the progress factor in the real-time activity progress information based on the project monitoring content, and judge the change parameter information within a preset range near the information segment corresponding to the progress factor, and judge the change parameter information corresponding to the progress factor; Step S3032: Based on the change parameter information corresponding to the progress factor, construct a progress parameter change model corresponding to the progress factor. The progress parameter change model is used to express the change of the progress parameter at different time nodes. Calculate the integral of the parameter change of the progress parameter change model over a continuous time period to obtain a continuous parameter change integral. If the continuous parameter change integral does not meet the preset standard, alarm the corresponding progress parameter change model. Step S3033: Denote the combination of the alarmed progress parameter change models as the key monitoring change model group, and analyze the key monitoring change model group to generate warning content.

7. The innovative project decision-making prompt method based on AI model analysis according to claim 6, characterized in that The method of analyzing the key monitoring change model group to generate warning content includes: Step S3034: Divide the progress parameter change model successively according to the equivalent parameter change interval, and determine the time length of each segmented model section, denoted as the time length for model comparison. Denote the combination of the time lengths for model comparison belonging to the same progress parameter change model as the time length group for comparison. Step S3035: Based on the equivalence of the progress parameter change models in the key monitoring change model group and the equivalence of the time length groups for comparison corresponding to each progress parameter change model, screen out the corresponding historical monitoring change model groups from a number of preset historical monitoring change model groups, and identify the warning content corresponding to the historical monitoring change model groups as the warning content of the key monitoring change model group.

8. An innovative project decision-making prompt system based on AI model analysis, characterized in that, Including: The first module is used to construct an innovation project template, and several project core labels are set for each innovation project template. The innovation project template includes the project activity type, and each project activity type is set with a project execution mapping tree. Each execution node of the project execution mapping tree is set with a project execution keyword and project monitoring content. The second module is used to perform real-time semantic understanding on the innovation project information, determine the project core labels and real-time activity progress information of the innovation project, compare the project core labels in the innovation project template, and determine the called innovation project template. The third module is used to determine the project execution keywords mapped in the project execution mapping tree based on the real-time activity progress information, and monitor the real-time activity progress information based on the project monitoring content corresponding to the project execution keywords to generate warning content.

Images