Artificial intelligence based visual data analysis application method

Abstract

The application provides an artificial intelligence-based visual data analysis application method and relates to the technical field of visual data mining. The career change recommendation model constructed by collecting career data of a first professional label and a set of alternative professional labels can give the career change matching degree between the first professional label and each alternative professional label, and the career change recommendation model can give the matching degree evaluation of the alternative professional label under adjacent position levels i-1 and i+1. A dynamic prediction model is constructed to generate the job transfer matching degree between the first professional label and each alternative professional label. The matching screening strategy can select the focus of the target user and implement a dynamic matching screening strategy, which has good flexibility in professional selection.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of visualizing data mining technology, in particular to an application method of visualizing data analysis based on artificial intelligence. BACKGROUND

[0002] Long Short-Term Memory (LSTM) in deep learning has shown great potential in dynamic data analysis due to its ability to process and predict time series data. This method can update and analyze large-scale career data sets in real time, helping individuals and enterprises make more data-supported decisions in career planning. However, in the existing environment, it is still a challenging technical task to show these complex analysis processes and results to users in an intuitive and understandable way.

[0003] In the prior art, a data analysis visualization method, system and device based on big data artificial intelligence, with publication number CN112883100A and name, includes establishing internet human resource big data talent demand index standard, and building a meta database based on it; according to the talent demand index standard established in step S1, collect data from the recruitment advertisements publicly available on the network, and use the big data platform to store the data; clean and desensitize the scraped web page data to form structured data of standard format; build a machine learning algorithm to analyze and mine data; and display the analysis results on an interactive real-time data visualization webpage.

[0004] The existing deficiencies are:

[0005] 1. Current target users usually lack a clear understanding of the prospects of the chosen career when making the first career choice; users' choices often rely on subjective judgments and lack a deep understanding of the future development trends of the chosen career; at the same time, they are not familiar with the related career options available for future career changes, resulting in high trial-and-error costs for individual career choices; existing matching and recommendation algorithms are usually based on static features, ignoring the dynamic nature of job changes and the path characteristics of individual career growth.

[0006] The above information disclosed in the above background section is only used to strengthen the understanding of the background of the present disclosure, and therefore it can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0007] The purpose of the present application is to provide an application method of visualizing data analysis based on artificial intelligence to solve the problems raised in the above background.

[0008] To achieve the above purpose, the present application provides the following technical solutions:

[0009] The application method of visual data analysis based on artificial intelligence comprises the following specific steps:

[0010] Step S1: Obtain career data corresponding to the first professional label, and obtain a first data set by data preprocessing on the career data;

[0011] The career data comprises, but is not limited to, an average working time index of different position levels, a salary level index, a position heat trend index and a required position skill demand index;

[0012] Step S2: Determine a set of candidate professional labels highly related to the first professional label;

[0013] Obtain career data corresponding to each candidate professional label in the set of candidate professional labels, and form a second data set by data preprocessing;

[0014] Step S3: Construct a professional change recommendation model, which is used to evaluate the professional change matching degree between the first professional label and each candidate professional label;

[0015] Step S4: Input the first data set and the second data set as professional development features into a long short-term memory network for training and verification to construct a dynamic prediction model, which generates the job transfer matching degree between the first professional label and each candidate professional label;

[0016] Step S5: Perform weight analysis on the professional change matching degree and the job transfer matching degree between the first professional label and each candidate professional label, display the weight analysis results by using a data visualization tool, and apply a corresponding matching screening strategy to the weight analysis results to screen at least three candidate professional labels meeting the professional development requirements from the set of candidate professional labels.

[0017] Compared with the prior art, the application has the beneficial effects that: by collecting the career data of the first professional label and the set of candidate professional labels, the constructed professional change recommendation model can give the professional change matching degree between the first professional label and each candidate professional label, and the professional change recommendation model can give the matching degree evaluation of the candidate professional label under the adjacent position levels i-1 and i+1, and by constructing a dynamic prediction model, the job transfer matching degree between the first professional label and each candidate professional label is generated; and the matching screening strategy can select the dynamic matching screening strategy according to the focus of the target user, and has good flexibility in professional selection. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 It is a schematic diagram of the overall method of the application. DETAILED DESCRIPTION

[0019] For purposes of the present invention, its technical solutions and advantages to be clearer and more apparent, the following further detailed description of the present invention with specific examples.

[0020] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood as the usual meaning understood by those skilled in the art to which the present invention belongs. The "first", "second" and similar words used in the present invention do not represent any order, quantity or importance, but are only used to distinguish different components. "Include" or "contain" and similar words mean that the elements or objects before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connected" or "connected" and similar words are not limited to physical or mechanical connection, but can include electrical connection, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to represent the relative positional relationship, when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0021] Example 1:

[0022] Please refer to Figure 1 The present invention provides a technical solution:

[0023] The visual data analysis application method based on artificial intelligence is used to provide alternative career choices for the first career label selected by the target user, and the specific steps include:

[0024] Step S1: Obtain career data corresponding to the first career label, and obtain a first data set by preprocessing the career data;

[0025] The career data includes but is not limited to average working years indicators, salary level indicators, job heat trend indicators and required job skill demand indicators of different job levels;

[0026] Further explanation: The acquisition of career data includes: obtaining the first career label selected by the target user from the career social platform, grabbing the public user profile of the first career label in the currently selected administrative region, and analyzing and processing the public user profile to generate career data;

[0027] The first career label is determined by the target user inputting in the career social platform or the career interest evaluation result. In this embodiment, the International Standard Classification of Occupations (ISCO-08) system is used for standardization of the career label to ensure the consistency and comparability of the career name;

[0028] Grabbing the public user profile of the first career label in the currently selected administrative region, the specific logic is:

[0029] A customized crawler is written using Scrapy framework to perform data scraping on selected data sources;

[0030] The scraped data fields include but are not limited to job title, job level, work experience, salary level, number of job postings, job posting frequency, and required skills;

[0031] During data scraping, geographic area filtering parameters are set to ensure that the collected data only comes from the user-selected administrative region, which includes the city or province to which it belongs.

[0032] Further, data preprocessing includes data cleaning of the selected data set, removing invalid and duplicate data, and desensitizing public user profiles to ensure data privacy and security, and storing the cleaned data in a structured database;

[0033] Data standardization operations include the following:

[0034] Job title standardization: unify similar job titles on different professional social platforms, such as "software engineer" and "software development engineer" to "software engineer".

[0035] Salary unification: convert monthly salary, annual salary, etc. in different units to annual salary and to a unified currency unit.

[0036] Specifically, scripts are written for string matching and unit conversion, and regular expressions are used to process text data.

[0037] Further, the skills required in the job description are segmented, stop words are removed, and key skill labels are identified and extracted;

[0038] Use natural language processing (NLP) techniques such as TF-IDF or Word2Vec for vector representation of skills; specifically use NLTK or spaCy for text processing and Scikit-learn for vectorization.

[0039] Integrate the pre-processed job average tenure indicator, salary level indicator, job heat trend indicator, and job skill demand indicator corresponding to the first career label into the first data set; and save the first data set in CSV or Parquet storage format;

[0040] Based on the first data set, the job average tenure indicator of the first career label k at job level i is defined as The larger the output value of job level i, the higher the job level;

[0041] The calculation formula is as follows:

[0042] ;

[0043] wherein, is the working years of the jth sample of the first professional label k under the position level i; M is the total number of samples; is the maximum working years of the first professional label k under the position level i; The output is limited in the interval (0, 1);

[0044] When The larger the output value is, the longer the average working years of the first professional label k under the position level i is; the more stable the position level i is;

[0045] Based on the first data set, the salary level index of the first professional label k under the position level i is defined as ; The calculation formula is as follows:

[0046] ;

[0047] is the salary of the jth sample of the first professional label k under the position level i;

[0048] is the maximum salary of the first professional label k under the position level i;

[0049] The output is limited in the interval (0, 1);

[0050] When The larger the output value is, the higher the salary of the first professional label k under the position level i is; the more prosperous the demand for the position level i is;

[0051] Based on the first data set, the position heat trend index of the first professional label k under the position level i is defined as ; The calculation formula is as follows:

[0052] ;

[0053] wherein, is the position publishing frequency of the jth sample of the first professional label k under the position level i;

[0054] is the historical maximum position publishing frequency of the first professional label k under the position level i, and the historical maximum position publishing frequency is determined based on the position publishing frequency in the last half year;

[0055] The output is limited to the interval (0, 1);

[0056] When The greater the output value, the higher the heat of the position level i of the first professional label k;

[0057] Based on the first data set, the position skill demand index of the first professional label k at the position level i is defined as ; The calculation formula is as follows:

[0058] ;

[0059] Wherein, is the number of samples in M samples of the first professional label k at the position level i that need to meet the required skill list; M is the total number of samples; The output is limited to the interval (0, 1);

[0060] The position skill demand index The corresponding required skill list is represented as ; Wherein h represents the index of the required skill, and H2 is the total number of required skill categories; The mapping of professional label and skill index is realized;

[0061] When The greater the output value, the higher the position skill demand of the first professional label k at the position level i.

[0062] Step S2: Determine a set of candidate professional labels highly related to the first professional label;

[0063] Obtain the career data corresponding to each candidate professional label in the set of candidate professional labels, and form a second data set after data preprocessing;

[0064] Further explanation: The determination of the set of candidate professional labels includes: obtaining the required skill list of the first professional label and each type of professional label in the professional database for difference analysis, so as to find out the set of candidate professional labels highly related to the first professional label;

[0065] The definition of "highly related" means that the required skill list of the candidate professional label has high coincidence with the required skill list of the first professional label, and the high coincidence is set to ≥ 70%; The specific logic is:

[0066] Based on the position skill demand index of the first professional label k at the position level i ; Determine the required skill list of the first professional label k at the position level i;

[0067] In the professional database, the professional labels of each type are screened under different position levels, and the professional labels containing at least 70% of the skills in the skill list required by the first professional label k under the position level i are obtained as a set of highly relevant candidate professional labels;

[0068] The working average length index, the salary level index, the position heat trend index and the position skill demand index of the career data corresponding to each candidate professional label in the candidate professional label set after preprocessing are integrated into a second data set;

[0069] Based on the second data set, the position skill demand index of each candidate professional label in the candidate professional label set under the position level i is marked as ; is the index of the professional label in the candidate professional label set, and K is the total number of candidate professional labels in the candidate professional label set, and i represents the index of the position level;

[0070] Based on the second data set, the position skill demand index of the candidate professional label under the position level i is defined as ;

[0071] Based on the second data set, the position skill demand index of the candidate professional label under the position level i is defined as ;

[0072] Based on the second data set, the position heat trend index of the candidate professional label under the position level i is defined as .

[0073] Step S3: Construct a career change recommendation model, which is used to evaluate the career change matching degree between the first professional label and each candidate professional label;

[0074] Further explanation: constructing a career change recommendation model, specifically including:

[0075] Based on the first data set and the second data set, the position skill demand indexes of the first professional label and each candidate professional label under the same position level combination are analyzed to obtain a first change matching coefficient;

[0076] Based on the first data set and the second data set, the position skill demand indexes of the first professional label and each candidate professional label under different span position level combinations are analyzed to obtain a second change matching coefficient;

[0077] Different span position level combinations refer to that the position level i of the first professional label k is combined with the position level i of the candidate professional label a position level i-1 or i+1 combination; i-1 and i+1 respectively represent an adjacent position level lower than the position level i and higher than the position level i;

[0078] comprehensive analysis of the first replacement matching coefficient and the second replacement matching coefficient to obtain a comprehensive matching index, the comprehensive matching index being used for comprehensive evaluation of the occupation replacement matching degree between the first occupation label and each candidate occupation label;

[0079] Further, the first replacement matching coefficient is represented by the following calculation formula:

[0080] ;

[0081] wherein, is the first occupation label k and the candidate occupation label The first replacement matching coefficient is obtained by data analysis of the position skill demand index under the position level i;

[0082] is the first occupation label k and the candidate occupation label The number of skill intersections under the position level i;

[0083] Increase : indicates that the candidate occupation label The skill matching degree under the same position level is improved;

[0084] is the total number of skills in the required skill list of the first occupation label k under the position level i;

[0085] Decrease : indicates that the required skills of the main occupation are reduced, and when other conditions remain unchanged, the matching coefficient is increased;

[0086] Based on the high coincidence between the required skill list of the candidate occupation label and the required skill list of the first occupation label, set The value range is ; is an interval representation;

[0087] When Tends to 1, it indicates that under the same position level i, the candidate occupation label The higher the coverage of the required skill list of the first occupation label k, the higher the occupation replacement matching degree, and the higher the recommendation priority;

[0088] The second replacement matching coefficient is represented by the following calculation formula:

[0089] ;

[0090] wherein, This refers to the relationship between the primary occupational label k and the alternative occupational labels under different job level combinations i-1 and i+1 with varying spans. The second replacement matching coefficient was obtained by analyzing the job skill requirement indicators.

[0091] The primary occupational label k is compared with the alternative occupational labels at job level i. Number of skill intersections at job level i-1;

[0092] The primary occupational label k is compared with the alternative occupational labels at job level i. The number of skill intersections at job level i+1;

[0093] Increase or This indicates that the skill match between the candidate occupation and the adjacent job level is improved.

[0094] The required skill lists for alternative job tags and the required skill lists for the primary job tag have a high degree of overlap. The range is ; It is an interval representation;

[0095] The second step in obtaining the matching coefficient is as follows:

[0096] Obtain the job skill demand index of the first occupation label k at job level i from the first dataset. .

[0097] Obtain candidate occupation labels from the second dataset. Job skill requirement indicators at job levels i-1 and i+1 and .

[0098] calculate and intersection .

[0099] calculate and intersection .

[0100] Add the quantities of the two intersections and divide by . To obtain the matching coefficient .

[0101] when The closer it is to 1, the more likely it is to be a candidate occupation label at adjacent job levels i-1 and i+1. The higher the coverage of the required skill list of the first professional label k, the higher the professional replacement matching degree, and the higher the recommendation priority.

[0102] The comprehensive matching index is a weighted integration of the first replacement matching coefficient and the second replacement matching coefficient. The comprehensive matching index is represented by the following calculation formula:

[0103] ;

[0104] wherein, is the comprehensive matching index, representing the professional replacement matching degree between the first professional label and each alternative professional label;

[0105] and is the weight coefficient of the corresponding replacement matching coefficient, used to allocate the weight of the first replacement matching and the second replacement matching in the comprehensive matching index; , and The values are all in the interval (0, 1). In this embodiment ;

[0106] Set the value range as ; is the interval representation;

[0107] When tends to 1, it means that under the same position level i or adjacent position levels i-1 and i+1, the alternative professional label The higher the coverage of the required skill list of the first professional label k, the higher the professional replacement matching degree, and the higher the recommendation priority.

[0108] The following example calculation is given in this embodiment:

[0109] The position skill requirement index of the first professional label k under the position level i=3 corresponds to the required skill list ;

[0110] The position skill requirement index of the alternative professional label under the position level i=3 corresponds to the required skill list ;

[0111] The position skill requirement index of the alternative professional label under the position level corresponds to the required skill list ;

[0112] The position skill requirement index of the alternative professional label under the position level corresponds to the required skill list ;

[0113] First replacement matching coefficient:

[0114] ;

[0115] Second replacement matching coefficient:

[0116]

[0117] Comprehensive matching index: set ;

[0118]

[0119] : alternative occupation label There is 60% skill matching under the same position level, indicating higher job transfer potential;

[0120] : alternative occupation label There is 40% skill matching under adjacent position levels, indicating certain job transfer flexibility.

[0121] : the comprehensive matching index is 0.5, indicating that the alternative occupation label has a moderate matching degree under multi-level position levels, and job transfer is suitable for consideration.

[0122] Step S4: input the first data set and the second data set as the occupation development feature set into the long short-term memory network for training and verification to construct a dynamic prediction model, and the dynamic prediction model generates the job transfer matching degree between the first occupation label and each alternative occupation label;

[0123] The first data set is represented as ;

[0124] Ensure that the number of samples of each index is consistent, providing consistency for subsequent model training;

[0125] The second data set is represented as ;

[0126] The and are taken as the occupation development feature set;

[0127] The long short-term memory (LSTM) model is selected for the following reasons:

[0128] LSTM is able to capture long-term dependencies in time series data, suitable for the time series changes of career development characteristics. Long short-term memory network (LSTM) is part of machine learning in the field of artificial intelligence;

[0129] For career development prediction, it can handle changing trends and volatility, and improve prediction accuracy.

[0130] A dynamic prediction model based on long short-term memory network is constructed, and the dynamic prediction model structure includes an input layer, multiple LSTM layers and an output layer;

[0131] The input layer receives a career development feature set composed of a first data set and a second data set;

[0132] After the career development feature set is processed by the LSTM layer, the final output layer generates the job transfer matching degree between the first career label and each alternative career label;

[0133] The following specific description is given in this embodiment:

[0134] Input layer: input dimension is and The combined feature set; the total number of features in this embodiment = 8;

[0135] LSTM layer: set 2 layers of LSTM, and the number of units in each layer is 64;

[0136] Fully connected layer: connected to an output layer, outputting the job transfer matching degree between the first career label and each alternative career label;

[0137] Use the training set (80% of the collected career development feature set for training) and the validation set (20% of the career development feature set for validation) to train the LSTM model, and the specific steps are:

[0138] Select the optimizer Adam and the loss function represented by the mean square error MSE;

[0139] Set the number of training rounds, initially set to 100 rounds, and dynamically adjust according to the validation loss when validating the results;

[0140] Use the validation set to evaluate the performance of the model, and ensure the prediction effect of the model by calculating the mean square error MS;

[0141] After the model training is completed, input the input data set and Into the trained LSTM model to generate the job transfer matching degree between the first career label and each alternative career label; specifically:

[0142] After the multi-layer LSTM processing, the last layer converts the learned feature information into an output corresponding to the alternative occupation label; through the activation function Softmax, the output result is converted into a value conforming to the probability distribution, which represents the transfer matching degree evaluation of each alternative occupation label to the first occupation label;

[0143] The numerical value in the dynamic prediction model output directly represents the transfer matching degree of the alternative occupation label to the first occupation label;

[0144] In this embodiment, when the dynamic prediction model outputs the matching degree of the alternative occupation label as 0.85 and the other as 0.50, the first alternative occupation label is more suitable for transfer from the first occupation;

[0145] It should be noted that the output layer generates the transfer matching degree between the first occupation label and each alternative occupation label, which specifically includes:

[0146] The output layer receives the data processed by the LSTM layer and converts it into a probability distribution form through the activation function Softmax;

[0147] The memory unit and gating mechanism inside the LSTM layer allow the LSTM to learn the features in the occupation development feature set that are relevant or influential in the past and future time;

[0148] The memory unit inside the LSTM model can learn the complex relationship between features, and through the weighting of input data, the model can not only evaluate the influence of each feature itself, but also understand the interaction between features;

[0149] Each LSTM layer receives input data and uses the learned weights in training to perform weighted calculation on the input.

[0150] The LSTM adjusts its memory state and hidden state at each time step to retain information useful for future prediction.

[0151] The last LSTM layer passes the processed data to the output layer to output the final prediction result, such as the matching degree score.

[0152] The output layer uses the activation function Softmax to convert the result into a probability distribution, reflecting the transfer matching degree of different occupation labels;

[0153] The output is the transfer matching degree of each alternative occupation label to the first occupation label k, which is denoted as The transfer matching degree of the occupation level i is denoted as ; the transfer matching degree The value range is between 0 and 1, and the transfer matching degree The greater the value, the higher the priority of the job change.

[0154] Step S5: Weight analysis is performed on the job change matching degree and the job change matching degree between the first career label and each candidate career label, the weight analysis results are displayed by using a data visualization tool, and a corresponding matching screening strategy is applied to the weight analysis results to screen at least three candidate career labels meeting the career development requirements from the candidate career label set.

[0155] Further explanation: The weight analysis is performed on the job change matching degree and the job change matching degree between the first career label and each candidate career label, and the specific logic is as follows:

[0156]

[0157] Among them, is the joint matching value of the first career label k and the candidate career label under the position level i;

[0158] and is the weight coefficient of the corresponding parameter, , and The values are in the interval (0, 1);

[0159] Let The effective value range of is the interval (0.35, 1];

[0160] The matching screening strategy specifically includes:

[0161] If , the candidate career label is marked as a high-priority job change career; , , and are the joint matching value, the comprehensive matching index, the second change matching coefficient and the judgment threshold of the job change matching degree, , , ; ;

[0162] When , it means that the number of job position level selections of the candidate career label is more under the adjacent position levels i-1 and i+1;

[0163] All candidate career labels are traversed, and at least three candidate career labels meeting the condition are preferentially screened, and if the number of screened labels is less than three, the The candidate professional label with the largest value is supplemented;

[0164] Further explanation: The weight analysis result of includes the following cases:

[0165] 1) When the target user has high requirements for the number of job position level selection and the professional replacement matching degree corresponding to the job position skill requirement index of the candidate professional label, set ; ;

[0166] The purpose of the setting is that, due to the high requirement for the job position skill requirement index, the number of job position level selection represented by needs to be considered first, and then the output value of the comprehensive matching index, which can highlight the number of job position level selection of the selected candidate professional label, and the flexibility is higher in the later career change. The purpose of setting the maximum is to improve the competitiveness of the candidate professional label in job change;

[0167] 2) When the target user has high requirements for the job average length index, the salary level index, and the job heat trend index of the candidate professional label, set ; ;

[0168] The purpose of the setting is that, due to the high requirement for the job average length index, the salary level index, and the job heat trend index, the job change matching degree represented by needs to be considered first, The prediction takes into account the job average length index, the salary level index, and the job heat trend index, so is considered as a high requirement; the technical effect of this setting can highlight that the candidate professional label can consider the selection requirements including the job average length index, the salary level index, and the job heat trend index, and the selected candidate professional label considers more aspects in the later career change;

[0169] 3) When the target user has high requirements for the job position skill requirement index, the job average length index, the salary level index, and the job heat trend index of the candidate professional label, set ; ;

[0170] , , and are determined by fuzzy analytic hierarchy process (FAHP), which will not be described here; and It is determined using the entropy weight method.

[0171] The purpose of this setting is to balance job skill requirements, average years of service, salary level, and job popularity trend indicators.

[0172] and The dynamic settings allow for optimization of matching results based on feedback from different job markets and user needs; this mechanism differs from static settings and can respond more flexibly to changes.

[0173] Use the matplotlib or seaborn libraries in Python to visualize the weight analysis results. The basic framework for data visualization is as follows:

[0174] Create a bar chart or pie chart to show the joint matching values ​​between different alternative occupation labels and the primary occupation label k. and The weighting analysis results will be used to visualize the final selected candidate occupational tags.

[0175] Data import: Use the pandas library to import the data into the working environment, ready for visualization;

[0176] Plotting functions: If using the matplotlib library: use plt.bar() or plt.pie() to create the corresponding charts;

[0177] If using the seaborn library: use seaborn.barplot() to create a bar chart;

[0178] Adjust details: Set the scales and labels for the x and y axes, as well as the position and style of the legend.

[0179] Chart display: Use the plt.show() command to display the chart on the screen for immediate viewing.

[0180] Save the chart: Use plt.savefig() to save the chart as an image file, including PNG or JPEG format, for embedding in reports or further sharing.

[0181] For candidate occupations with a joint match value higher than 0.7, different colors are used to highlight occupations with higher recommendation priority.

[0182] Example 2:

[0183] Further illustrated on the basis of Embodiment One, in order to verify the effectiveness of the artificial intelligence-based visual data analysis application method, "data analyst" is selected as the first professional label k, and the set of candidate professional labels includes "data scientist", "business analyst", "machine learning engineer", "big data engineer", "BI analyst", and "statistician". In the implementation process, first, the career data of each professional label under the position level i is collected and sorted, including the average working years, salary level, position heat trend, and required position skill demand indicators. The specific steps are as follows:

[0184] 1.1) Data acquisition and preprocessing:

[0185] Through industry reports, recruitment websites, and professional survey data, the relevant data of "data analyst" and candidate professional labels at different position levels (junior, intermediate, and senior) are obtained. Data preprocessing includes data cleaning, missing value processing, and standardization processing to ensure the consistency and comparability of the data. In this embodiment, for "data scientist" at the intermediate position, the average working years are 3 years, the salary level is 15 million RMB / year, the position heat trend shows an upward trend, and the required skills include Python, machine learning, and data visualization.

[0186] 1.2) Construction of career change recommendation model:

[0187] Based on the collected career data, a career change recommendation model is constructed. The model calculates the comprehensive matching index between "data analyst" and each candidate professional label to determine potential job transfer opportunities. In this embodiment, by calculating the matching degree of "data analyst" and "data scientist" at the intermediate position, the career change matching degree is 0.85.

[0188] 1.3) Training and verification of dynamic prediction model:

[0189] The first data set and the second data set are input into the long short-term memory network (LSTM) for training and verification to construct a dynamic prediction model to generate the job transfer matching degree . In this embodiment, after training, the predicted value of the job transfer matching degree of "data analyst" and "business analyst" at the senior position is 0.78, indicating that this job transfer opportunity has a high priority.

[0190] 1.4) Weight analysis and matching screening:

[0191] The career change matching degree and the job transfer matching degree are analyzed for weights, the joint matching value is calculated, and the matching screening strategy is applied to screen out candidate professional labels that meet the requirements of career development. The weight coefficients and The entropy weight method is used to ensure that the comprehensive matching value is in the effective interval (0.35, 1]. In this embodiment, for the high requirement position skill demand index and the occupation replacement matching degree, b1=0.6 and b2=0.4 are set, and the joint matching value of the "machine learning engineer" is calculated as 0.92, which meets the screening standard.

[0192] 1.5) Visual result display:

[0193] A bar chart is generated using the Matplotlib library to visually display the joint matching values between each alternative occupation label and "data analyst". Different matching degrees are distinguished by color, and high matching value occupation labels are highlighted to assist users in making scientific career transition decisions.

[0194] Through the above implementation steps, not only the application process of the visual data analysis method based on artificial intelligence in career matching is demonstrated, but also the advantages of the method in matching accuracy and decision support are proved through specific data. The following gives the experimental data table:

[0195] Table 1: Job transfer position matching results:

[0196]

[0197] The table is explained as follows:

[0198] The above table shows the matching indicators of "data analyst" and six alternative occupation labels at the intermediate to senior position levels. The comprehensive matching index and the job transfer matching degree reflect the suitability of occupation replacement and job transfer, and the joint matching value comprehensively evaluates the priority of job transfer through weight analysis. Experimental data show that the visual data analysis method based on artificial intelligence can effectively identify alternative occupation labels with high matching degree, and assist users in making scientific decisions in the process of career development. In this embodiment, "machine learning engineer" has a higher comprehensive matching index and job transfer matching degree, which is significantly better than other alternative occupations, reflecting the superiority and innovation of the method in career matching;

[0199] In the case of weight setting b1=0.6, b2=0.4, this method gives priority to the diversity of job transfer positions. Once a certain alternative occupation label, such as "machine learning engineer", shows high adaptability at multiple position levels, it will perform more outstandingly in the screening process;

[0200] In addition, the calculation of job transfer matching degree directly considers dynamic market demand and career development trends, so for positions with extensive skill requirements and upward trends, the system identifies higher job transfer flexibility. This not only provides individuals with more career development path choices, but also enhances their competitiveness in an uncertain job market, ensuring that they can be transferred to positions that meet individual and market needs. This flexibility is particularly important in rapidly changing industries, where users can find the most suitable development path among more job level choices.

[0201] It should be noted that all the calculation formulas in this application file use regression analysis in machine learning algorithms, including but not limited to, to deeply analyze the collected relevant parameters and identify their natural trends and mutual relationships. Professional software such as Python's Scikit-learn library or R language is used to automatically generate mathematical models that match the data. Then, through methods such as cross-validation, the performance of the model is objectively evaluated, and combined with continuous feedback and optimization, to ensure that the created formula truly reflects the inherent laws of the data, thereby ensuring its effectiveness and accuracy, and ensuring that the calculation process complies with the constraints of natural laws, rather than being based on artificially set rules.

[0202] The technical solutions of the present application can be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), FLASH, hard disk or optical disk, etc. The computer software product includes a number of instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods of various embodiments of the present application.

[0203] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a list of executable instructions for implementing logic functions, which can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute the instructions) or in conjunction with these instruction execution systems, devices or apparatus. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transport programs for use by an instruction execution system, device or apparatus or in conjunction with these instruction execution systems, devices or apparatus.

[0204] It should be noted that the above examples are only used to illustrate the technical solutions of the present application but not limit the present application. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced without departing from the spirit and scope of the present application, and all should be covered in the scope of the claims of the present application.

[0205] It should be noted that the above examples are only used to illustrate the technical solutions of the present application but not limit the present application. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced without departing from the spirit and scope of the present application, and all should be covered in the scope of the claims of the present application.

Claims

1. An artificial intelligence based visual data analysis application method for providing alternative career choices for a first career label selected by a target user, characterized in that, The specific steps include: Step S1: Obtain career data corresponding to the first professional label, and obtain a first data set by preprocessing the career data; The career data includes average working years indicators, salary level indicators, job heat trend indicators, and required job skill demand indicators of different position levels; Step S2: Determine a set of candidate professional labels highly related to the first professional label; Obtain career data corresponding to each candidate professional label in the set of candidate professional labels, and form a second data set by preprocessing the data; Step S3: Build a career change recommendation model, which is used to evaluate the career change matching degree between the first professional label and each candidate professional label; Specifically, it includes: Based on the first data set and the second data set, the job skill demand indicators of the first professional label and each candidate professional label are analyzed under the same position level combination to obtain a first replacement matching coefficient; Based on the first data set and the second data set, the job skill demand indicators of the first professional label and each candidate professional label are analyzed under different span position level combinations to obtain a second replacement matching coefficient; Different span combinations of job levels refer to: combining job level i of first career label k with job level i-1 or i+1 of alternative career label ; i-1 and i+1 represent adjacent job levels lower than job level i and higher than job level i, respectively. Comprehensive analysis of the first replacement matching coefficient and the second replacement matching coefficient obtains a comprehensive matching index, which is used to comprehensively evaluate the career change matching degree between the first professional label and each candidate professional label; Step S4: Input the first data set and the second data set as a career development feature set into a long short-term memory network for training and verification to build a dynamic prediction model, which generates the job transfer matching degree between the first professional label and each candidate professional label; A dynamic prediction model based on a long short-term memory network is built, and the dynamic prediction model structure includes an input layer, multiple LSTM layers, and an output layer; The input layer receives a career development feature set composed of the first data set and the second data set; After processing the career development feature set through the LSTM layer, the output layer finally generates the job transfer matching degree between the first professional label and each candidate professional label; Step S5: Weight analysis is performed on the career change matching degree and the job transfer matching degree between the first professional label and each candidate professional label, and the weight analysis results are displayed using a data visualization tool. Apply the corresponding matching filtering strategy to these weight analysis results to filter out at least three candidate professional labels that meet the career development requirements from the set of candidate professional labels.

2. The artificial intelligence based visual data analysis application method according to claim 1, characterized in that: The acquisition of career data includes: obtaining the first professional label selected by the target user from the professional social platform, crawling the public user data of the first professional label in the current selected administrative region, and analyzing and processing the public user data to generate career data; The working average years indicators, salary level indicators, job heat trend indicators, and job skill demand indicators of the first professional label corresponding to the preprocessed career data are integrated into the first data set; Based on the first data set, the job average tenure indicator of the first professional label k at the position level i is defined as The greater the position level i output value, the higher the position level. Based on the first data set, the salary level indicator of the first professional label k at the position level i is defined as ; Based on the first data set, a job heat trend indicator of the first professional label k at the position level i is defined as ; When The greater the output value, the higher the heat of the position level i representing the first professional label k. Based on the first data set, a job skill demand indicator of a first professional label k at a job level i is defined as ; The job skill requirement index The corresponding required skill list is represented as ; wherein h represents the index of the required skill, and H2 is the total number of required skill categories. When The greater the output value, the higher the position skill demand of the first professional label k under the position level i.

3. The artificial intelligence based visual data analysis application method of claim 2, wherein: The determination of the alternative occupation label set comprises: obtaining the first occupation label and the required skill list of each type of occupation label in the occupation database for difference analysis to find a set of alternative occupation labels highly related to the first occupation label; The definition of "highly related" means that the required skill list of the alternative occupation label has high coincidence with the required skill list of the first occupation label, and the high coincidence is set as the number of coincident skills ≥ 70%; The working average length index, the salary level index, the position heat trend index and the position skill demand index of the pre-processed career data of each alternative occupation label in the alternative occupation label set are integrated into a second data set; based on the second data set, an index of the job skill demand of each alternative career label in the set of alternative career labels at job level i is marked as ; is an index of the career label in the set of alternative career labels, and K is the total number of alternative career labels in the set of alternative career labels; i represents the index of the job level. based on the second data set, an alternative professional label The average tenure indicator for a job at position rank i is defined as ; Based on the second data set, alternative professional labels are identified The salary level indicator under the job grade i is defined as ; Based on the second data set, alternative professional labels are determined The position heat trend indicator under the position rank i is defined as .

4. The artificial intelligence based visual data analysis application method of claim 3, wherein: The first replacement matching coefficient is represented by the following calculation formula: ； wherein, is a first replacement matching coefficient for the first professional label k and the alternative professional label The position skill requirement index under the position level i is analyzed to obtain a first replacement matching coefficient. is the first professional label k and the alternative professional label is the number of skill intersections under the position rank i; is the total number of skills of the list of required skills of the first professional label k under the position rank i; Based on the high coincidence of the required skill list of the alternative occupation label with the required skill list of the first occupation label, it is determined The value domain is ; is an interval representation; When The closer to 1, the higher the recommended priority, indicating that the candidate job label The higher the coverage of the required skill list of the first job label k, the higher the job replacement matching degree, and the higher the recommended priority. The second replacement matching coefficient is represented by the following calculation formula: ； wherein, is a second replacement matching coefficient obtained by data analysis on the first career label k and the alternative career label 's job skill requirement index under different span of position level combination i-1 and i+1. is the number of first professional labels k at position level i that intersect with the alternative professional label the number of skill intersections at position level i-1; is the number of skill intersections of the first professional label k at position level i with the alternative professional label at position level i+1; based on the required skill list of the alternative career tag having a high degree of overlap with the required skill list of the first career tag, setting a value range ; is an interval representation; When The closer to 1, the higher the job change match degree, the higher the recommendation priority, indicates that the candidate job label The higher the coverage of the required skill list of the first job label k, the higher the job change match degree, the higher the recommendation priority.

5. The artificial intelligence based visual data analysis application method of claim 4, wherein: The comprehensive matching index is represented by the following calculation formula: ； wherein, is a comprehensive matching index, and the output result represents the occupation change matching degree between the first occupation label and each alternative occupation label; and are weight coefficients of the respective replacement matching coefficients, , and all take values in the interval (0, 1); Set The value range is ; is an interval representation; When The closer to 1, the higher the job change match degree, the higher the recommendation priority, indicates that the candidate job label The higher the coverage of the required skill list of the first job label k, the higher the job change match degree, the higher the recommendation priority.

6. The artificial intelligence based visual data analysis application method of claim 5, wherein: characterizing the first data set as ; characterizing the second data set as ; will be described below. and as a career development feature set; The transfer matching degree between the first occupation label and each alternative occupation label is generated through the output layer, specifically including: The output layer receives the data processed by the LSTM layer and converts it into a probability distribution form through the activation function Softmax; The output is each alternative career label The first career label k is matched with the alternative career label The job transfer matching degree under the position level i is denoted as The job transfer matching degree The value range is between 0 and 1, and the job transfer matching degree The greater the value, the higher the priority of the job transfer.

7. The artificial intelligence-based visual data analysis application method of claim 6, wherein: The weight analysis of the occupation replacement matching degree and the transfer matching degree between the first occupation label and each alternative occupation label is performed, and the specific logic is: ； wherein, is the first professional label k and the alternative professional label is the joint match value under the position rank i; and are weight coefficients of the respective parameters, , and all take values in the interval (0, 1); Set The valid value range of is the interval (0.35, 1]; The setting of the matching screening strategy specifically includes: If , the alternative occupation label is marked as a high-priority job transfer occupation; , , and are respectively the joint matching value, the comprehensive matching index, the second replacement matching coefficient and the judgment threshold of the job transfer matching degree, , , ; ; At the time, the representation of the alternative career label has more job level selection number in the adjacent job levels i-1 and i+1; Iterate through all candidate occupation tags and prioritize selecting those that match at least three. If the number of alternative occupational tags for the given conditions is less than three, then select those that meet the criteria. Filter from the alternative occupation tags The candidate occupation tags with the highest values ​​are used to supplement the list.

8. The artificial intelligence based visual data analysis application method of claim 7, wherein: and The weight analysis results of and include the following cases: 1) When the target user only selects a number of job position levels corresponding to the job position level of the job skill requirement index of the alternative occupation label and requires a high degree of occupation change matching, set ; ; 2) When the target user's requirements for the average working years indicator, the salary level indicator, and the job heat trend indicator of the alternative occupation label are high, set ; ; 3) When the target user has high requirements for the position skill requirement index, the average working years index, the salary level index, and the position heat trend index of the alternative occupation label, set ; ; The data visualization of the weight analysis result is performed using the matplotlib or seaborn library in Python.

Images