Employment post matching method and system based on data analysis

Abstract

The invention discloses an employment post matching method and system based on data analysis, relates to the technical field of employment post matching, constructs a space-time elastic matching system adaptive to a flexible employment scene, breaks through the limitation of traditional static time presetting through a sliding window and thermodynamic diagram analysis, and improves the matching efficiency. Stability and temporary changes of real available time periods of users are accurately captured, and the method is especially suitable for time period fluctuation requirements of housewives and other groups caused by family affairs; based on time granularity division and core time period marking of industry characteristics, efficient utilization of fragmentary time periods such as household service and the like is met, the requirement for continuity of equipment operation posts is guaranteed, and intelligent adaptation of macroscopic requirements and microcosmic time units is achieved; meanwhile, interaction modes are automatically switched according to physiological features of special employment groups, and the exposure priority of core competitive posts is improved while matching fairness is guaranteed through an advantage dimension weight adjustment mechanism.

Description

Technical Field

[0001] The present invention relates to the technical field of job matching, and in particular to a job matching method and system based on data analysis. Background Art

[0002] With the rapid development of the gig economy and shared employment models, flexible employment platforms are gradually introducing intelligent matching algorithms to improve the efficiency of matching people with jobs. Intermittent workers such as housewives and childcare workers have become important participants in this model, and their employment periods are fragmented and dynamic.

[0003] For this group of people, current matching systems use reinforcement learning-based task allocation models, such as DeepQ-Network, to predict users' potential availability times based on historical order data. Alternatively, they employ multi-agent collaborative algorithms to break down conflicting tasks in overlapping time periods into subtask sequences. However, these solutions rely on pre-set time windows to capture temporary time changes, such as shortened available time periods caused by emergencies, and dynamic adjustments rely solely on delayed human feedback. Multi-task splitting strategies are limited by fixed time granularity, typically measured in units of one hour, making them difficult to adapt to short, fragmented time periods, such as 15 minutes.

[0004] Although some solutions introduce graph neural networks to construct time dependencies, or encode time periods into vector spaces for similarity matching, their underlying assumption is still that the user's available time periods are continuous and regular. For intermittent employment scenarios, the actual order acceptance rate is significantly lower than the theoretical matching value. Summary of the Invention

[0005] In view of the above existing problems, the present invention is proposed.

[0006] The present invention provides a job matching method and system based on data analysis to solve the problems that existing flexible employment platforms are difficult to effectively deal with the dynamic time conflicts of intermittent employees, low micro-time utilization and lack of adaptation for special groups.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, an embodiment of the present invention provides a job position matching method based on data analysis, which includes:

[0009] Step S1: Obtain the time distribution data and real-time location information from the user's historical order records and build a time period preference prediction model. The time period preference prediction model extracts order records of the same time period within three consecutive working cycles through a sliding time window and marks the stable available time period.

[0010] Step S2: parse the spatiotemporal constraints in the job requirement text to generate multi-dimensional requirement features including time slot sequence, geographic location, and skill tags;

[0011] Step S3: Deconstructing continuous job requirements into discrete time units through a task splitting algorithm and dynamically aligning them with the user's time slot preference prediction results. The dynamic alignment includes calculating the optimal travel path time for each time unit based on the road network topology relationship between the user's real-time location and the task location, and dynamically adjusting the start time of the time slot sequence.

[0012] Step S4: Combine the blockchain evidence information of the user's skill tags and the job requirement characteristics to calculate the multi-dimensional matching degree and generate a recommendation sequence.

[0013] As a preferred solution of the job position matching method based on data analysis described in the present invention, the construction of the time period preference prediction model includes:

[0014] Extract the effective working hours, task intervals and cancellation records from the user's historical order records;

[0015] Generate a time period activity heat map based on time distribution density to identify stable and temporary user availability periods.

[0016] Establish a time period availability probability matrix and dynamically update the impact weights of sudden time period change events.

[0017] As a preferred embodiment of the data analysis-based job matching method of the present invention, in step S1, for the order records within the sliding time window, a time period activity heat map is generated through the following process, and the stable available time period and the temporary available time period are identified based on the heat map, including:

[0018] In the last three working cycles, the time axis is divided into L time period particles, and the number of orders received on the i-th particle in the k-th window is recorded as x k,i , k=1,2,3,i=1,…,L, based on this definition of order density:

[0019]

[0020] Among them, p k,i represents the order density of the kth window on particle i, x k,i represents the number of orders received by the k-th window on the time period particle i, j represents the traversal index, and L represents the total number of time period particles;

[0021] The density mean of the three windows is used as the particle activity, and the activity standard deviation is calculated:

[0022]

[0023] Among them, H i represents the average activity of the i-th particle, σ i represents the activity fluctuation of particle i;

[0024] Perform statistics on the global activity distribution and obtain the average value and standard deviation:

[0025]

[0026] Among them, μ H is the overall average activity level, σ H is the overall fluctuation of activity;

[0027] Then use the activity dispersion coefficient to determine the core threshold:

[0028] w=1+λ·CV H ,α=μ H +wσ H ,

[0029] Among them, CV H represents the activity dispersion coefficient, λ is the sensitivity adjustment coefficient, w is the dynamic weight coefficient, and α is the activity determination threshold;

[0030] To distinguish volatility situations, a volatility threshold is defined:

[0031]

[0032] Among them, median({σ i}) is the median volatility of all particles, δ is the volatility threshold adjustment coefficient,

[0033] Based on this, the time periods are divided into:

[0034]

[0035] Among them, S stable To stabilize the available set, S temp A temporarily available collection.

[0036] As a preferred solution of the job position matching method based on data analysis described in the present invention, the task splitting algorithm includes:

[0037] Determine the time granularity based on the industry standards corresponding to the job requirement type, and divide the job requirement duration into combinable time units. For example, housekeeping services use a 15-minute granularity, and logistics and distribution use a 30-minute granularity.

[0038] When it is detected that the job requirement text contains preset keywords, the corresponding time period is marked as an inseparable core period. The preset keywords include continuous operation and equipment preheating.

[0039] Establish a time unit dependency diagram, marking the logical constraints between the indivisible core time period and the adjustable auxiliary time period;

[0040] Through the conflict resolution mechanism based on time window overlap detection, the priorities of overlapping tasks of multiple users are assigned.

[0041] As a preferred solution of the job position matching method based on data analysis described in the present invention, the processing of the blockchain evidence information includes:

[0042] Call the lightweight certificate feature extraction model to parse the anti-counterfeiting mark of the user skill label card;

[0043] Build a certificate template feature library and automatically verify the validity of certificates through similarity comparison;

[0044] Calculate the semantic relevance between the verified skill tags and the job requirement tags;

[0045] The calculation of the multi-dimensional matching degree includes:

[0046] Set dynamic weight coefficients for time period matching, geographical location matching, and skill matching respectively;

[0047] When it is detected that the user belongs to a special employment group, the weight coefficient of their advantage dimension is automatically increased;

[0048] The geographic location matching degree is dynamically corrected based on real-time traffic data.

[0049] As a preferred solution of the job position matching method based on data analysis described in the present invention, in step S4, when it is detected that the user belongs to a special employment group, the weight coefficient of the user's advantage dimension is automatically increased, including:

[0050] In the multi-dimensional matching dimension set D = {d1, d2, d3}, let m d Represents the matching score of dimension d, and the base weight coefficient is w d ,∑ d∈D w d =1, where D represents the matching dimension set, including time period, geography, and skills, m d represents the original matching score of dimension d, w d represents the baseline weight coefficient of dimension d;

[0051] Identify the strengths dimension:

[0052]

[0053] Among them, d adv indicates the dimension with the highest score;

[0054] Detect user special group indication

[0055] If the user belongs to a special employment group, then G = 1, otherwise G = 0,

[0056] If G=1, adjust the weights of each dimension:

[0057]

[0058] Among them, β∈(0,1) is the lifting coefficient, w′ d is the adjusted weight, if G = 0, then w′ d =w d ;

[0059] Among them, d adv represents the advantage dimension with the highest matching degree, G represents the indicator of special employment groups, β represents the weight improvement coefficient, and w′ d represents the adjusted weight of dimension d after detecting special groups.

[0060] In a second aspect, the present invention provides a job matching system based on data analysis, comprising:

[0061] The data collection module is used to obtain the spatiotemporal behavior data uploaded by the user's mobile terminal and the skill tags stored on the blockchain;

[0062] Model building module with built-in time period preference prediction model and task splitting algorithm engine;

[0063] Matching engine, including multi-dimensional feature alignment unit and dynamic weight calculation unit;

[0064] The output module generates a recommendation list with timing conflict warning signs and a visual matching path diagram.

[0065] As a preferred solution of the job matching system based on data analysis described in the present invention, the spatiotemporal behavior data includes the GPS trajectory point set of the user's mobile terminal, the Wi-Fi positioning hotspot switching frequency and the Bluetooth beacon contact record, and the sensitive location information is desensitized through differential privacy technology.

[0066] As a preferred solution of the job position matching system based on data analysis described in the present invention, the matching engine further includes:

[0067] The certificate verification submodule communicates with the blockchain node and calls the feature extraction model;

[0068] Special group adaptation unit, which automatically switches to non-voice interaction mode when hearing-impaired / visual-impaired labels are recognized;

[0069] The real-time correction unit updates the parameters of the time period preference prediction model based on user feedback data.

[0070] As a preferred solution of the job matching system based on data analysis described in the present invention, the non-voice interaction mode includes: when a hearing-impaired label is detected, the voice notification is automatically converted to a vibration reminder mode, and the contrast of the text prompt box is enhanced in the visual interface.

[0071] The beneficial effects of the present invention are as follows: the present invention constructs a spatiotemporal elastic matching system that is adapted to flexible employment scenarios. Through sliding window and heat map analysis, it breaks through the limitations of traditional static time presets, accurately captures the stability and temporary changes of users' real available time periods, and is especially adapted to the time period fluctuations caused by family affairs for groups such as housewives; based on the time granularity division and core time period marking of industry characteristics, it not only meets the efficient utilization of fragmented time periods such as housekeeping services, but also ensures the continuity requirements of equipment operation positions, and realizes the intelligent adaptation of macro needs and micro time units; at the same time, it automatically switches the interaction mode according to the physiological characteristics of special employment groups (such as hearing impairment), and through the advantage dimension weight adjustment mechanism, it improves the exposure priority of its core competitiveness positions while ensuring matching fairness.

[0072] The combination of blockchain evidence storage and lightweight certificate verification ensures the authenticity and timeliness of skill labels, effectively solving the matching distortion problem caused by false skill labels in traditional platforms. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0074] Figure 1 This is a flow chart of the job position matching method based on data analysis in Example 1.

[0075] Figure 2 This is a schematic diagram of the framework of the job matching system based on data analysis in Example 1. DETAILED DESCRIPTION

[0076] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0077] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0078] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0079] Example 1, with reference to Figure 1 and Figure 2 This embodiment provides a job position matching method based on data analysis, comprising the following steps:

[0080] Step S1: Obtain the time distribution data and real-time location information from the user's historical order records and build a time period preference prediction model. The time period preference prediction model extracts order records of the same time period within three consecutive working cycles through a sliding time window and marks the stable available time period.

[0081] The construction of the time period preference prediction model includes:

[0082] Extract the effective working hours, task intervals and cancellation records from the user's historical order records;

[0083] Generate a time period activity heat map based on time distribution density to identify stable and temporary user availability periods.

[0084] Establish a time period availability probability matrix and dynamically update the impact weight of sudden time period change events;

[0085] In step S1, for the order records within the sliding time window, a time period activity heat map is generated through the following process, and the stable available time period and temporary available time period are identified based on the heat map, including:

[0086] In the last three working cycles, the time axis is divided into L time period particles, and the number of orders received on the i-th particle in the k-th window is recorded as x k,i , k=1,2,3,i=1,…,L, based on this definition of order density:

[0087]

[0088] Among them, p k,i represents the order density of the kth window on particle i, x k,irepresents the number of orders received by the k-th window on the time period particle i, j represents the traversal index, and L represents the total number of time period particles;

[0089] The density mean of the three windows is used as the particle activity, and the activity standard deviation is calculated:

[0090]

[0091] Among them, H i represents the average activity of the i-th particle, σ i represents the activity fluctuation of particle i;

[0092] Perform statistics on the global activity distribution and obtain the average value and standard deviation:

[0093]

[0094] Among them, μ H is the overall average activity level, σ H is the overall fluctuation of activity;

[0095] Then use the activity dispersion coefficient to determine the core threshold:

[0096] w=1+λ·CV H ,α=μ H +wσ H ,

[0097] Among them, CV H represents the activity dispersion coefficient, λ is the sensitivity adjustment coefficient, which is automatically adjusted according to the user activity distribution, w is the dynamic weight coefficient, and α is the activity determination threshold;

[0098] To distinguish volatility situations, a volatility threshold is defined:

[0099]

[0100] Among them, median({σ i}) is the median volatility of all particles, δ is the volatility threshold adjustment coefficient,

[0101] Based on this, the time periods are divided into:

[0102]

[0103] Among them, S stable To stabilize the available set, S temp For a temporarily available collection,

[0104] Specifically, this process uses density normalization to unify the measurement of order-taking behaviors in different work cycles. Combining the dual indicators of mean and volatility, it can not only highlight long-term high-activity periods, but also eliminate the interference of occasional peaks in judgment. The dynamic weight coefficient is calculated from the activity dispersion coefficient, and can automatically adjust the threshold sensitivity according to the fluctuation of user behavior. It is applicable to both highly stable users and highly volatile users. The median volatility is used as an auxiliary threshold, and combined with the activity threshold, it realizes the effective identification of temporarily available time periods. The generated heat map can intuitively present the activity distribution of each time period. This method takes into account both stability and sensitivity and is universal for diversified order-taking models.

[0105] Step S2: parse the spatiotemporal constraints in the job requirement text to generate multi-dimensional requirement features including time slot sequence, geographic location, and skill tags;

[0106] Step S3: Deconstruct the continuous job requirements into discrete time units through a task splitting algorithm and dynamically align them with the user's time slot preference prediction results. Dynamic alignment includes: calculating the optimal travel path time for each time unit based on the road network topology relationship between the user's real-time location and the task location, and dynamically adjusting the start time of the time slot sequence;

[0107] The task splitting algorithm includes:

[0108] Determine the time granularity based on the industry standards corresponding to the job requirement type, and divide the job requirement duration into combinable time units. For example, housekeeping services use a 15-minute granularity, and logistics and distribution use a 30-minute granularity.

[0109] When it is detected that the job requirement text contains preset keywords, the corresponding time period will be marked as an inseparable core period. The preset keywords include continuous operation and equipment preheating.

[0110] Create a time unit dependency diagram, marking the logical constraints between the inseparable core time periods and the adjustable auxiliary time periods;

[0111] Priority is assigned to overlapping tasks of multiple users through a conflict resolution mechanism based on time window overlap detection.

[0112] Step S4: Calculate the multi-dimensional matching degree and generate a recommendation sequence based on the blockchain evidence of the user's skill tags and job requirements.

[0113] The processing of blockchain evidence information includes:

[0114] Call the lightweight certificate feature extraction model to parse the anti-counterfeiting mark of the user skill label card;

[0115] Build a certificate template feature library and automatically verify the validity of certificates through similarity comparison;

[0116] Calculate the semantic relevance between the verified skill tags and the job requirement tags;

[0117] The calculation of multi-dimensional matching includes:

[0118] Set dynamic weight coefficients for time period matching, geographical location matching, and skill matching respectively;

[0119] When it is detected that the user belongs to a special employment group, the weight coefficient of their advantage dimension is automatically increased;

[0120] Dynamically revise the geographic location matching degree based on real-time traffic data;

[0121] In step S4, when it is detected that the user belongs to a special employment group, the weight coefficient of the user's advantage dimension is automatically increased, including:

[0122] In the multi-dimensional matching dimension set D = {d1, d2, d3}, let m d Represents the matching score of dimension d, and the base weight coefficient is w d ,∑ d∈D w d =1, where D represents the matching dimension set, including time period, geography, and skills, m d represents the original matching score of dimension d, w d represents the baseline weight coefficient of dimension d;

[0123] Identify the strengths dimension:

[0124]

[0125] Among them, d adv indicates the dimension with the highest score;

[0126] Detect user special group indication

[0127] If the user belongs to a special employment group, then G = 1, otherwise G = 0,

[0128] If G=1, adjust the weights of each dimension:

[0129]

[0130] Among them, β∈(0,1) is the lifting coefficient, w′ d is the adjusted weight, if G = 0, then w′ d =w d ;

[0131] Among them, d adv represents the advantage dimension with the highest matching degree, G represents the indicator of special employment groups, β represents the weight improvement coefficient, and w ′d represents the adjusted weight of dimension d after detecting a special group;

[0132] Specifically, this step balances fairness and efficiency by increasing the weight of the advantage dimension with the highest matching score. After increasing the weight of the advantage dimension, special employment groups will obtain a higher overall matching degree in their areas of expertise, thereby improving the accuracy and satisfaction of recommended positions. The use of normalization adjustment can ensure that the total weight is always 1, and will not destroy the original multi-dimensional matching structure. The improvement coefficient β can be flexibly set according to the policy needs of different groups to achieve controllable rights and interests tilt. This mechanism can be easily expanded to more dimensions in actual systems, and can also be linked with real-time feedback data to dynamically optimize the weight strategy.

[0133] This embodiment also provides a job matching system based on data analysis, including:

[0134] The data collection module is used to obtain the spatiotemporal behavior data uploaded by the user's mobile terminal and the skill tags stored on the blockchain;

[0135] Model building module with built-in time period preference prediction model and task splitting algorithm engine;

[0136] Matching engine, including multi-dimensional feature alignment unit and dynamic weight calculation unit;

[0137] Output module, generates a recommendation list with timing conflict warning signs and a visual matching path diagram;

[0138] Spatiotemporal behavior data includes the GPS trajectory points of the user's mobile terminal, the switching frequency of Wi-Fi positioning hotspots, and the contact records of Bluetooth beacons. Sensitive location information is desensitized using differential privacy technology.

[0139] The matching engine also includes:

[0140] The certificate verification submodule communicates with the blockchain node and calls the feature extraction model;

[0141] Special group adaptation unit, which automatically switches to non-voice interaction mode when hearing-impaired / visual-impaired labels are recognized;

[0142] A real-time correction unit that updates the parameters of the time period preference prediction model based on user feedback data;

[0143] Non-voice interaction modes include: automatically converting voice notifications to vibration reminder mode when a hearing-impaired label is detected, and enhancing the contrast of text prompt boxes in the visual interface.

[0144] 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 the present invention. Although the present invention 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 invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A job position matching method based on data analysis, characterized in that: include, Step S1: Obtain the time distribution data and real-time location information from the user's historical order records and build a time period preference prediction model. The time period preference prediction model extracts order records of the same time period within three consecutive working cycles through a sliding time window and marks the stable available time period. Step S2: parse the spatiotemporal constraints in the job requirement text to generate multi-dimensional requirement features including time slot sequence, geographic location, and skill tags; Step S3: Deconstruct the continuous job requirements into discrete time units through the task splitting algorithm and dynamically align them with the user's time period preference prediction results; The dynamic alignment includes: calculating the optimal travel path time of each time unit based on the road network topology relationship between the user's real-time location and the task location, and dynamically adjusting the starting time of the time slot sequence; Step S4: Combine the blockchain evidence information of the user's skill tags and the job requirement characteristics to calculate the multi-dimensional matching degree and generate a recommendation sequence.

2. The job position matching method based on data analysis according to claim 1, characterized in that: The construction of the time period preference prediction model includes: Extract the effective working hours, task intervals and cancellation records from the user's historical order records; Generate a time period activity heat map based on time distribution density to identify stable and temporary user availability periods. Establish a time period availability probability matrix and dynamically update the impact weights of sudden time period change events.

3. The job position matching method based on data analysis according to claim 2, characterized in that: In step S1, for the order records within the sliding time window, a time period activity heat map is generated through the following process, and the stable available time period and temporary available time period are identified based on the heat map, including: In the last three working cycles, the time axis is divided into L time period particles, and the number of orders received on the i-th particle in the k-th window is recorded as x k,i , k=1,2,3,i=1,…,L, based on this definition of order density: Among them, p k,i represents the order density of the kth window on particle i, x k,i represents the number of orders received by the k-th window on the time period particle i, j represents the traversal index, and L represents the total number of time period particles; The density mean of the three windows is used as the particle activity, and the activity standard deviation is calculated: Among them, H i represents the average activity of the i-th particle, σ i represents the activity fluctuation of particle i; Perform statistics on the global activity distribution and obtain the average value and standard deviation: Among them, μ H is the overall average activity level, σ H is the overall fluctuation of activity; Then use the activity dispersion coefficient to determine the core threshold: w=1+λ·CV H ,a=m H +wσ H , Among them, CV H represents the activity dispersion coefficient, λ is the sensitivity adjustment coefficient, w is the dynamic weight coefficient, and α is the activity determination threshold; To distinguish volatility situations, a volatility threshold is defined: Among them, median({σ i }) is the median volatility of all particles, δ is the volatility threshold adjustment coefficient, Based on this, the time periods are divided into: Among them, S stable To stabilize the available set, S temp A temporarily available collection.

4. The job position matching method based on data analysis according to claim 1, characterized in that: The task splitting algorithm includes: Determine the time granularity based on the industry standards corresponding to the job requirement type, and divide the job requirement duration into combinable time units. For example, housekeeping services use a 15-minute granularity, and logistics and distribution use a 30-minute granularity. When it is detected that the job requirement text contains preset keywords, the corresponding time period is marked as an inseparable core period. The preset keywords include continuous operation and equipment preheating. Establish a time unit dependency diagram, marking the logical constraints between the indivisible core time period and the adjustable auxiliary time period; Through the conflict resolution mechanism based on time window overlap detection, the priorities of overlapping tasks of multiple users are assigned.

5. The job position matching method based on data analysis according to claim 1, characterized in that: The processing of blockchain evidence information includes: Call the lightweight certificate feature extraction model to parse the anti-counterfeiting mark of the user skill label card; Build a certificate template feature library and automatically verify the validity of certificates through similarity comparison; Calculate the semantic relevance between the verified skill tags and the job requirement tags; The calculation of the multi-dimensional matching degree includes: Set dynamic weight coefficients for time period matching, geographical location matching, and skill matching respectively; When it is detected that the user belongs to a special employment group, the weight coefficient of their advantage dimension is automatically increased; The geographic location matching degree is dynamically corrected based on real-time traffic data.

6. The job position matching method based on data analysis according to claim 5, characterized in that: In step S4, when it is detected that the user belongs to a special employment group, the weight coefficient of the user's advantage dimension is automatically increased, including: In the multi-dimensional matching dimension set D = {d1, d2, d3}, let m d Represents the matching score of dimension d, and the base weight coefficient is w d ,∑ d∈D w d =1, where D represents the matching dimension set, including time period, geography, and skills, m d represents the original matching score of dimension d, w d represents the baseline weight coefficient of dimension d; Identify the strengths dimension: Among them, d adv indicates the dimension with the highest score; Detect user special group indication If the user belongs to a special employment group, then G = 1, otherwise G = 0, If G=1, adjust the weights of each dimension: Among them, β∈(0,1) is the lifting coefficient, w′ d is the adjusted weight, if G = 0, then w′ d =w d ; Among them, d adv represents the advantage dimension with the highest matching degree, G represents the indicator of special employment groups, β represents the weight improvement coefficient, and w ′ d represents the adjusted weight of dimension d after detecting special groups.

7. A job position matching system based on data analysis, based on a job position matching method based on data analysis according to any one of claims 1 to 6, characterized in that: include: The data collection module is used to obtain the spatiotemporal behavior data uploaded by the user's mobile terminal and the skill tags stored on the blockchain; Model building module with built-in time period preference prediction model and task splitting algorithm engine; Matching engine, including multi-dimensional feature alignment unit and dynamic weight calculation unit; The output module generates a recommendation list with timing conflict warning signs and a visual matching path diagram.

8. The job position matching system based on data analysis according to claim 7, characterized in that: The spatiotemporal behavior data includes the GPS trajectory point set of the user's mobile terminal, the Wi-Fi positioning hotspot switching frequency and the Bluetooth beacon contact record, and the sensitive location information is desensitized using differential privacy technology.

9. The job position matching system based on data analysis according to claim 7, characterized in that: The matching engine also includes: The certificate verification submodule communicates with the blockchain node and calls the feature extraction model; Special group adaptation unit, which automatically switches to non-voice interaction mode when hearing-impaired / visual-impaired labels are recognized; The real-time correction unit updates the parameters of the time period preference prediction model based on user feedback data.

10. The job position matching system based on data analysis according to claim 9, characterized in that: The non-voice interaction mode includes: when a hearing-impaired label is detected, automatically converting the voice notification into a vibration reminder mode, and enhancing the contrast of the text prompt box in the visual interface.

Images