Guarantee product recommendation system and method based on big data

Abstract

The invention discloses a guarantee product recommendation system and method based on big data, relates to the technical field of big data analysis, and realizes refined management of doctor-seeing data by finely dividing hospital departments, constructing a department doctor-seeing fluctuation analysis set and combining quarterly division of years. And extracting fluctuation data of a target year and a target quarter, predicting the number of patients in combination with the current year, and accurately predicting the number of patients in each quarter by means of a time sequence prediction model. And screening out a recommendation department with the most doctor seeing number, analyzing age data of the recommendation department to construct a distribution diagram, and determining an age interval with the most proportion as a guarantee product recommendation object. According to the method, department business interference is avoided, the change rule of the number of patients is clearly presented, the prediction accuracy of the number of patients is improved, the promotion of guarantee products is more targeted, efficient connection of medical resources and the guarantee products is realized, and the overall efficiency of medical guarantee services is improved.

Description

Technical Field

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

[0002] In the field of medical insurance product recommendations, existing technologies usually perform risk modeling based on static historical incidence rates or basic user health data. However, there are significant technical limitations when responding to dynamic scenarios of public health emergencies and subsequent mass preventive interventions. When a disease causes a sudden increase in the rate of visits due to seasonal epidemics, the subsequent incidence rate deviates significantly from historical patterns. This makes it difficult to accurately match the actual health risks and protection needs of the public when recommending corresponding medical insurance products to them, and it is impossible to formulate a scientific, reasonable and targeted medical insurance product recommendation strategy based on the dynamic changes in disease risks. Summary of the invention

[0003] The purpose of the present invention is to provide a security product recommendation system and method based on big data to solve the problems raised in the prior art.

[0004] To achieve the above object, the present invention provides the following technical solution: a method for recommending insurance products based on big data, the method for recommending insurance products comprising the following steps: Step S1, divide the departments of the hospital, select a single department as the target department, and divide all the departments of the hospital into multiple single departments; and divide each year into quarters; obtain the medical record data of the target department in each year after the quarterly division to construct a department medical visit fluctuation analysis set; Step S1-1, obtaining the department type of the hospital, dividing the departments of the same type into a single department, and selecting a single department as the target department; Step S1-2: The quarter division adopts the equal-length time period definition method, dividing each year into N continuous and non-overlapping equal-length time periods according to natural months, each equal-length time period includes a complete natural month with a fixed number of days; each year contains N identical quarters; Step S1-3, obtain the target department's medical record data in each quarter of each year through the hospital information system to construct a department medical visit fluctuation analysis set, wherein the medical record data is represented by the number of patients corresponding to the target department; the department medical visit fluctuation analysis set is used to store the medical record data of the target department's number of patients in each quarter of each year.

[0005] The single department is specifically represented by a single independent department unit with a specific diagnosis and treatment scope and professional division of labor under the hospital structure; the target departments include internal medicine, surgery, pediatrics and ophthalmology, etc. Dividing hospital departments into single departments can accurately focus on specific diagnosis and treatment fields, making the analysis more targeted and avoiding cross-interference of different departments' business on the accuracy and reliability of data. The quarterly division of equal-length periods ensures the consistency and comparability of data in each time period, which helps to clearly present the regular changes in the number of patients in the time dimension.

[0006] Step S2, extracting the year with the largest fluctuation of the target department's medical record data as the target year through the department medical record fluctuation analysis set analysis, and using the medical record data of the target year as the target year fluctuation data; extracting the quarter with the largest fluctuation of the medical record data as the target quarter according to the target year fluctuation data analysis, and using the medical record data of the target quarter as the target quarter fluctuation data; Step S2-1, extracting the target department's visit record data according to the department visit fluctuation analysis set analysis, taking the year with the largest number of patients in each year as the target year, and taking the visit record data of the target year as the target year fluctuation data; Step S2-2, traverse and read the target year fluctuation data to obtain the medical record data of each quarter in the target year. The quarter with the largest number of patients is the quarter with the largest fluctuation in the medical record data, which is taken as the target quarter; the number of patients in the target quarter is taken as the target quarter fluctuation data.

[0007] Extracting the target year and target quarter from the department visit fluctuation analysis set can provide valuable key time node information for medical data analysis. Determining the target year and focusing on the year with the largest number of patients can quickly lock in the peak year of department business volume, helping hospitals to clearly understand the changing trend of department business scale between different years.

[0008] Step S3, according to the target year fluctuation data and the target quarter fluctuation data, combined with the medical record data of each quarter of the current year, obtain the fluctuation change value of the medical record data of the target department in the current year, and use the department medical record fluctuation analysis set combined with the medical record data fluctuation change value to predict the number of patients in the quarter corresponding to the target quarter in the current year; Step S3-1, analyzing the quarter in the target year fluctuation data that is the same as the quarter of the current year, and obtaining the medical record data in the current year that is the same as the month of the target quarter; Step S3-2, according to the department visit fluctuation analysis set, calculate the average number of patients visiting the target department in the target quarter as the target quarterly standard value, and calculate the average number of patients visiting the target department in each quarter as the quarterly average value; Step S3-3, counting the number of quarters that have passed in the current year as the number of quarters that have passed, and counting the medical records of the number of quarters that have passed as the current cumulative number of medical visits; and counting the medical records in the target year according to the number of quarters that have passed as the reference cumulative number of medical visits; Step S3-4, using the current cumulative number of patients divided by the product of the number of past quarters and the quarterly average value to obtain a first ratio, multiplying the target quarterly standard value by the first ratio to obtain the seasonal number of patients corresponding to the target season in the current year as the first predicted value; Step S3-5, dividing the current cumulative number of patients by the reference cumulative number of patients to obtain a second ratio, multiplying the target quarterly fluctuation data by the second ratio to obtain the seasonal number of patients corresponding to the target season in the current year as the second predicted value; Step S3-6, according to the historical predicted values ​​of the number of patients visiting the target department, analyze and extract the first predicted value weight and the second predicted value weight used in calculating the predicted value of the number of patients visiting each time, and construct a weight set; Step S3-7, calculate the prediction error variance E according to the weight set, and dynamically allocate the first prediction value weight and the second prediction value weight through the smoothing factor a, and obtain the predicted value of the number of patients in the season corresponding to the target season in the current year through weighted data fusion; The calculation formula for the predicted value of the number of patients obtained by weighted data fusion is as follows: Y=w1×y1+w2×y2; Where, Y represents the predicted value of the number of patients obtained by weighted data fusion; w1 represents the weight of the first predicted value; w2 represents the weight of the second predicted value; y1 represents the first predicted value; y2 represents the second predicted value; The error variance is calculated using the following formula: ; In the formula, E i is the error variance of the i-th prediction value; n is the number of i-th prediction values ​​in the weight set; e new，i It is expressed as the error of the actual number of patients when predicting the i-th predicted value; e i Expressed as the error of the i-th prediction value; The variance update calculation uses the following formula: ; In the formula, E1 is represented by the error variance of the first predicted value; a is represented by the smoothing factor, and its value range is 0<a<1; e1 is represented by the error predicted by the first predicted value, which is obtained by subtracting the predicted number of patients from the actual number of patients; u1 is represented by the average error predicted by the first predicted value; ← is represented by the assignment operation; specifically, when any value of the error variance of the first predicted value, the error predicted by the first predicted value, and the average error predicted by the first predicted value changes, the calculation result on the right side of ← will also change; when the calculation parameter on the right side of ← changes, the changed value is used to perform calculation according to the given calculation method, and the calculation result is assigned to E1 on the left side of ← to complete the assignment operation.

[0009] The calculation formula for dynamically adjusting the weight of the first prediction value and the weight of the second prediction value according to the weight set is as follows: ; Where E2 is the error variance of the second prediction value and satisfies the condition: w2=1-w1; By comparing and analyzing the data of the same quarter of the target year and the current year, we can closely follow the actual business situation and dig out the associated changes in the trend of medical visits within the year. Calculate the standard value of the target quarter and the quarterly average to provide a basis for the standardized analysis of the data, making the forecast more scientific and stable. By statistically analyzing the current quarterly data and comparing it with the corresponding data of the target year, constructing the first ratio and the second ratio, the number of patients can be estimated from different angles, which can comprehensively consider the impact of historical data fluctuations and current business progress on the future. Finally, the weighted fusion of the two predicted values ​​is adopted, which takes into account the advantages of different prediction methods and greatly improves the accuracy of the prediction of the number of patients.

[0010] Step S4, analyzing the result of the patient number prediction to obtain the number of patients in each target department, and selecting the target department with the largest number of patients as the recommended department according to the number of patients in each target department; The number of patients visiting the target department in the same quarter in different years is input into the time series prediction model to predict the number of patients in each quarter. The total number of patients visiting the target department in the current year is obtained by adding the current cumulative number of patients and the predicted number of patients. The total number of patients visiting each target department in the current year is sorted, and the target department with the largest number of patients is selected as the recommended department.

[0011] The time series prediction model is calculated using the following formula: ; Where P t It is the predicted value of the number of patients in quarter t; a is the constant term; n is the order of the autoregressive part, which is to use the data of the past n quarters to predict the current value; Y iExpressed as the autoregression coefficient, specifically the number of patients in the previous quarter P t-i The number of patients in the current quarter P t The degree of influence; m represents the order of the moving average part, specifically using the error terms of the past m quarters to predict the current value; R j Expressed as a moving average coefficient; F t-j Expressed as the error term in quarter tj; By inputting the number of patients visiting the target department in the same quarter in different years into the time series forecasting model, we can fully explore the time series rules in the historical data, and thus accurately predict the number of patients visiting each quarter. Combined with the current cumulative number of patients, the calculation of the total number of patients visiting the target department in the current year is more comprehensive and accurate, reflecting the combination of current actual business conditions and future trends. Sorting the total number of patients visiting each target department and selecting the department with the most patients as the recommended department will help hospitals clearly identify the departments with the largest business volume and the most concentrated demand.

[0012] Step S5: Analyze and extract the age data of the recommended departments, construct an age distribution chart based on the age data, and select the age range with the largest proportion as the recommended object for the insurance product recommendation based on the age distribution chart; Step S5-1, obtaining the age of patients in the recommended departments through the hospital information system, constructing a horizontal axis based on the age data, and constructing an age distribution chart based on the number of patients corresponding to different ages as the vertical axis; Step S5-2: for the constructed age distribution graph, take the set age range as the unit, count the number of patients in each age range, calculate the proportion of the number of patients in each age range to the total number of patients, and screen out the age range with the highest proportion by comparing the proportion of each age range. The people in this age range will be recommended as the objects for insurance product recommendation.

[0013] By accurately obtaining the age data of the recommended departments through the hospital information system and constructing an age distribution chart, the overall age distribution of the patients in the department can be intuitively and clearly displayed. Statistical analysis is performed based on the set age ranges, and the proportion of patients in each range is calculated, which helps to accurately locate the age group with the most concentrated departmental business. Screening out the people in the age range with the highest proportion as the recommended objects for protection products can make the promotion and design of protection products more targeted. On the one hand, medical insurance product suppliers can optimize product terms and protection scope based on this precise positioning, so that they are more in line with the actual medical needs of the target population and enhance the practicality and attractiveness of the products; on the other hand, hospitals and related institutions can use this result to provide more suitable protection product information for patients in specific age groups, improve patients' acceptance and utilization rate of protection products, thereby effectively improving the overall efficiency of medical insurance services and realizing efficient docking and precise delivery of medical resources and protection products.

[0014] Further, a big data-based insurance product recommendation system includes a department and time division module, a key data extraction module, a patient number prediction module, a recommended department screening module and an insurance object determination module; The department and time division module is used to subdivide the hospital departments into single departments and complete the division of each year and quarter; the key data extraction module is used to extract the target year and target quarter data from the department's visit fluctuation data; the number of patients predicted module is used to predict the number of patients in the target quarter based on historical and current data; the recommended department screening module is used to screen out the department with the largest number of patients based on the number of patients predicted results; the protection object determination module is used to determine the recommended object of the protection product through the age data of the recommended department; The output end of the department and time division module is electrically connected to the input end of the key data extraction module; the output end of the key data extraction module is electrically connected to the input end of the patient number prediction module; the output end of the patient number prediction module is electrically connected to the input end of the recommended department screening module; the output end of the recommended department screening module is electrically connected to the input end of the security object determination module; The department and time division module includes a department division unit and a quarter division unit; the department division unit is responsible for dividing all departments of the hospital into a single department and selecting a target department; the quarter division unit is used to divide each year into quarters using an equal-length time period definition method; The key data extraction module includes a target year extraction unit and a target quarter extraction unit; the target year extraction unit is used to find the year with the largest number of patients from the department visit fluctuation analysis set; the target quarter extraction unit is used to determine the quarter with the largest number of patients in the target year data; The patient number prediction module includes a basic data statistics unit and a prediction value calculation unit; the basic data statistics unit is used to obtain the corresponding quarterly data of the current year and calculate various average values ​​and cumulative number of patients; the prediction value calculation unit is used to obtain the first prediction value and the second patient number prediction value by different methods and merge them; The recommended department screening module includes a patient number forecasting and summarizing unit and a department screening unit; the patient number forecasting and summarizing unit is used to use a time series model to predict and summarize the number of patients in each quarter; the department screening unit is used to sort the number of patients in each department and select the recommended department; The protection object determination module includes an age distribution drawing unit and a protection object screening unit; the age distribution drawing unit is used to obtain the age data of recommended departments to construct a distribution map; the protection object screening unit is used to divide age intervals and screen the people in the interval with the highest proportion as the recommended objects for protection product recommendations.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention realizes the refined management and analysis of medical data by dividing hospital departments and constructing department visit fluctuation analysis sets, combined with the quarterly division of the year. This can not only accurately focus on specific diagnosis and treatment fields and avoid cross-interference of data accuracy by different departments, but also clearly present the regular changes in the number of patients in the time dimension, providing a comprehensive and reliable data foundation for subsequent in-depth analysis, and effectively making up for the shortcomings of existing technologies in data processing.

[0016] 2. The present invention predicts the number of patients by comprehensively using the fluctuation data of the target year and the target quarter, combined with the medical records of the current year. A variety of methods are used to calculate the predicted values ​​and weighted fusion, fully exploring the impact of historical data and current business progress on the future, and using the time series prediction model to accurately predict the number of patients in each quarter, greatly improving the accuracy of the number of patients predicted. This makes it possible to more accurately grasp the medical needs of different departments when responding to dynamic changes in disease risks, providing strong support for the formulation of scientific and reasonable insurance product recommendation strategies, and overcoming the defect that the existing technology is difficult to accurately match insurance needs in dynamic scenarios.

[0017] 3. The present invention constructs an age distribution diagram by analyzing the age data of the recommended departments and selects the age range with the largest proportion as the recommended objects of insurance products. This measure makes the promotion and design of insurance products more targeted. Medical insurance product suppliers can optimize products according to the actual medical needs of the target population. Hospitals and related institutions can also provide more suitable insurance product information for patients of specific age groups, thereby improving patients' acceptance and utilization rate of insurance products, improving the overall efficiency of medical insurance services, and realizing efficient docking and precise delivery of medical resources and insurance products. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic diagram of a process of recommending insurance products based on big data according to the present invention; Figure 2 The present invention is a structural diagram of a security product recommendation system based on big data. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] Embodiment 1: Figure 1As shown, the present invention provides a technical solution, a security product recommendation method based on big data, and the security product recommendation method includes the following steps: Step S1, divide the departments of the hospital, select a single department as the target department, and divide all the departments of the hospital into multiple single departments; and divide each year into quarters; obtain the medical record data of the target department in each year after the quarterly division to construct a department medical visit fluctuation analysis set; Step S1-1, obtaining the department type of the hospital, dividing the departments of the same type into a single department, and selecting a single department as the target department; Step S1-2: The quarter division adopts the equal-length time period definition method, dividing each year into N continuous and non-overlapping equal-length time periods according to natural months, each equal-length time period includes a complete natural month with a fixed number of days; each year contains N identical quarters; Step S1-3, obtain the target department's medical record data in each quarter of each year through the hospital information system to construct a department medical visit fluctuation analysis set, wherein the medical record data is represented by the number of patients corresponding to the target department; the department medical visit fluctuation analysis set is used to store the medical record data of the target department's number of patients in each quarter of each year.

[0021] In the specific implementation, there are many departments in the hospital and the business is complicated. Subdividing them into a single department can enable the subsequent analysis to focus on specific professional fields. The equal-length period definition method is used to divide the quarters to ensure the consistency and comparability of each time period on the time scale. The use of hospital information systems to obtain medical record data can effectively collect the number of patients in each target department in each year and quarter because of its comprehensive data and timely updates.

[0022] Step S2, extracting the year with the largest fluctuation of the target department's medical record data as the target year through the department medical record fluctuation analysis set analysis, and using the medical record data of the target year as the target year fluctuation data; extracting the quarter with the largest fluctuation of the medical record data as the target quarter according to the target year fluctuation data analysis, and using the medical record data of the target quarter as the target quarter fluctuation data; Step S2-1, extracting the target department's visit record data according to the department visit fluctuation analysis set analysis, taking the year with the largest number of patients in each year as the target year, and taking the visit record data of the target year as the target year fluctuation data; Step S2-2, traverse and read the target year fluctuation data to obtain the medical record data of each quarter in the target year, the quarter with the largest number of patients is the quarter with the largest fluctuation in the medical record data, and is taken as the target quarter; the number of patients in the target quarter is taken as the target quarter fluctuation data; In the specific implementation, the year with the largest number of patients is selected as the target year from the department visit fluctuation analysis set, because the year with the peak number of patients can reflect the activity and change trend of the department's business. In the target year data, the quarter with the largest number of patients is further determined as the target quarter, based on the importance of the business concentration period to the analysis. This layered screening method can accurately locate the key time nodes of the department's business.

[0023] Step S3, according to the target year fluctuation data and the target quarter fluctuation data, combined with the medical record data of each quarter of the current year, obtain the fluctuation change value of the medical record data of the target department in the current year, and use the department medical record fluctuation analysis set combined with the medical record data fluctuation change value to predict the number of patients in the quarter corresponding to the target quarter in the current year; Step S3-1, analyzing the quarter in the target year fluctuation data that is the same as the quarter of the current year, and obtaining the medical record data in the current year that is the same as the month of the target quarter; Step S3-2, according to the department visit fluctuation analysis set, calculate the average number of patients visiting the target department in the target quarter as the target quarterly standard value, and calculate the average number of patients visiting the target department in each quarter as the quarterly average value; Step S3-3, counting the number of quarters that have passed in the current year as the number of quarters that have passed, and counting the medical records of the number of quarters that have passed as the current cumulative number of medical visits; and counting the medical records in the target year according to the number of quarters that have passed as the reference cumulative number of medical visits; Step S3-4, using the current cumulative number of patients divided by the product of the number of past quarters and the quarterly average value to obtain a first ratio, multiplying the target quarterly standard value by the first ratio to obtain the seasonal number of patients corresponding to the target season in the current year as the first predicted value; Step S3-5, dividing the current cumulative number of patients by the reference cumulative number of patients to obtain a second ratio, multiplying the target quarterly fluctuation data by the second ratio to obtain the seasonal number of patients corresponding to the target season in the current year as the second predicted value; Step S3-6, according to the historical predicted values ​​of the number of patients visiting the target department, analyze and extract the first predicted value weight and the second predicted value weight used in calculating the predicted value of the number of patients visiting each time, and construct a weight set; Step S3-7, calculate the prediction error variance E according to the weight set, and dynamically allocate the first prediction value weight and the second prediction value weight through the smoothing factor a, and obtain the predicted value of the number of patients in the season corresponding to the target season in the current year through weighted data fusion; In specific implementation, comparing the data of the target year and the same quarter of the current year can closely combine with the actual business scenario and dig out the correlation of the annual medical treatment trend. Calculate the target quarter standard value and quarterly average value to provide a basis for data standardization, so that data from different years and quarters are comparable. By constructing the first ratio and the second ratio, the number of patients can be estimated from different angles (historical data fluctuations and current business progress).

[0024] Step S4, analyzing the result of the patient number prediction to obtain the number of patients in each target department, and selecting the target department with the largest number of patients as the recommended department according to the number of patients in each target department; The number of patients visiting the target department in the same quarter in different years is input into the time series prediction model to predict the number of patients in each quarter. The total number of patients visiting the target department in the current year is obtained by adding the current cumulative number of patients and the predicted number of patients. The total number of patients visiting each target department in the current year is sorted, and the target department with the largest number of patients is selected as the recommended department.

[0025] In the specific implementation, the number of patients in the same quarter of different years of the target department is input into the time series forecasting model, which uses the model's deep analysis ability of time series data to explore the potential rules and trends in it, so as to predict the number of patients in future quarters. Combined with the current cumulative number of patients, it not only takes into account the long-term trend reflected by historical data, but also takes into account the current actual business progress, making the calculated total number of patients in the target department in the current year more comprehensive and accurate.

[0026] Step S5: Analyze and extract the age data of the recommended departments, construct an age distribution chart based on the age data, and select the age range with the largest proportion as the recommended object for the insurance product recommendation based on the age distribution chart; Step S5-1, obtaining the age of patients in the recommended departments through the hospital information system, constructing a horizontal axis based on the age data, and constructing an age distribution chart based on the number of patients corresponding to different ages as the vertical axis; Step S5-2: for the constructed age distribution graph, the number of patients in each age range is counted, and the proportion of the number of patients in each age range to the total number of patients is calculated. By comparing the proportion values ​​of each age range, the age range with the highest proportion is selected, and the people in this age range are recommended as the objects for the recommendation of the insurance product; In specific implementation, the age data of the recommended departments is obtained through the hospital information system and an age distribution chart is constructed to intuitively display the age distribution of the patients. The proportion of patients is counted in the set age range, aiming to accurately locate the age group with the most concentrated department business, and provide a target population basis for product recommendation.

[0027] Embodiment 2, as Figure 2 As shown, the present invention provides a big data-based insurance product recommendation system, which includes a department and time division module, a key data extraction module, a patient number prediction module, a recommended department screening module, and an insurance object determination module; The department and time division module is used to subdivide the hospital departments into single departments and complete the division of each year and quarter; the key data extraction module is used to extract the target year and target quarter data from the department's visit fluctuation data; the number of patients predicted module is used to predict the number of patients in the target quarter based on historical and current data; the recommended department screening module is used to screen out the department with the largest number of patients based on the number of patients predicted results; the protection object determination module is used to determine the recommended object of the protection product through the age data of the recommended department; The output end of the department and time division module is electrically connected to the input end of the key data extraction module; the output end of the key data extraction module is electrically connected to the input end of the patient number prediction module; the output end of the patient number prediction module is electrically connected to the input end of the recommended department screening module; the output end of the recommended department screening module is electrically connected to the input end of the security object determination module; The department and time division module includes a department division unit and a quarter division unit; the department division unit is responsible for dividing all departments of the hospital into a single department and selecting a target department; the quarter division unit is used to divide each year into quarters using an equal-length time period definition method; The key data extraction module includes a target year extraction unit and a target quarter extraction unit; the target year extraction unit is used to find the year with the largest number of patients from the department visit fluctuation analysis set; the target quarter extraction unit is used to determine the quarter with the largest number of patients in the target year data; The patient number prediction module includes a basic data statistics unit and a prediction value calculation unit; the basic data statistics unit is used to obtain the corresponding quarterly data of the current year and calculate various average values ​​and cumulative number of patients; the prediction value calculation unit is used to obtain the first prediction value and the second patient number prediction value by different methods and merge them; The recommended department screening module includes a patient number forecasting and summarizing unit and a department screening unit; the patient number forecasting and summarizing unit is used to use a time series model to predict and summarize the number of patients in each quarter; the department screening unit is used to sort the number of patients in each department and select the recommended department; The protection object determination module includes an age distribution drawing unit and a protection object screening unit; the age distribution drawing unit is used to obtain the age data of recommended departments to construct a distribution map; the protection object screening unit is used to divide age intervals and screen the people in the interval with the highest proportion as the recommended objects for protection product recommendations.

[0028] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A method for recommending insurance products based on big data, characterized in that: The guarantee product recommendation method comprises the following steps: Step S1, divide the departments of the hospital into multiple single departments, select a single department as the target department; and divide each year into quarters; obtain the medical record data of the target department in each year after the division, and construct a department medical visit fluctuation analysis set; Step S2, extracting the year with the largest fluctuation of the target department's medical record data as the target year through the department medical record fluctuation analysis set analysis, and using the medical record data of the target year as the target year fluctuation data; extracting the quarter with the largest fluctuation of the medical record data as the target quarter according to the target year fluctuation data analysis, and using the medical record data of the target quarter as the target quarter fluctuation data; Step S3, according to the target year fluctuation data and the target quarter fluctuation data, combined with the medical record data of each quarter of the current year, obtain the fluctuation change value of the medical record data of the target department in the current year, and use the department medical record fluctuation analysis set combined with the medical record data fluctuation change value to predict the number of patients in the quarter corresponding to the target quarter in the current year; Step S4, analyzing the result of the patient number prediction to obtain the number of patients in each target department, and selecting the target department with the largest number of patients as the recommended department according to the number of patients in each target department; Step S5: Analyze and extract the age data of the recommended departments, construct an age distribution chart based on the age data, and select the age range with the largest proportion as the recommended object for insurance product recommendation based on the age distribution chart.

2. The method for recommending insurance products based on big data according to claim 1, characterized in that: The specific steps of step S1 are as follows: Step S1-1, obtaining the department type of the hospital, dividing the departments of the same type into a single department, and selecting a single department as the target department; Step S1-2: The quarter division adopts the equal-length time period definition method, dividing each year into N continuous and non-overlapping equal-length time periods according to natural months, each equal-length time period includes a complete natural month with a fixed number of days; each year contains N identical quarters; Step S1-3, obtaining the target department's medical record data in each quarter of each year through the hospital information system, and constructing a department medical visit fluctuation analysis set, wherein the medical record data is represented by the number of patients corresponding to the target department.

3. The method for recommending insurance products based on big data according to claim 2, characterized in that: The specific steps of step S2 are as follows: Step S2-1, extracting the target department's visit record data according to the department visit fluctuation analysis set analysis, taking the year with the largest number of patients in each year as the target year, and taking the visit record data of the target year as the target year fluctuation data; Step S2-2, traverse and read the target year fluctuation data to obtain the medical record data of each quarter in the target year. The quarter with the largest number of patients is the quarter with the largest fluctuation in the medical record data, which is taken as the target quarter; the number of patients in the target quarter is taken as the target quarter fluctuation data.

4. The method for recommending insurance products based on big data according to claim 3, characterized in that: The specific steps of step S3 are as follows: Step S3-1, obtaining the medical record data of the same month as the target quarter in the current year, and analyzing the fluctuation of the medical record data of the quarter of the current year; Step S3-2, according to the department visit fluctuation analysis set, calculate the average number of patients visiting the target department in the target quarter as the target quarterly standard value, and calculate the average number of patients visiting the target department in each quarter as the quarterly average value; Step S3-3, counting the number of quarters that have passed in the current year as the number of passed quarters, and counting the medical record data of the number of passed quarters as the current cumulative number of medical visits; and counting the medical record data in the target year according to the number of passed quarters as the reference cumulative number of medical visits.

5. The method for recommending insurance products based on big data according to claim 4, characterized in that: In step S3, it also includes: Step S3-4, using the current cumulative number of patients divided by the product of the number of past quarters and the quarterly average value to obtain a first ratio, multiplying the target quarterly standard value by the first ratio to obtain the seasonal number of patients corresponding to the target season in the current year as the first predicted value; Step S3-5, dividing the current cumulative number of patients by the reference cumulative number of patients to obtain a second ratio, multiplying the target quarterly fluctuation data by the second ratio to obtain the seasonal number of patients corresponding to the target season in the current year as the second predicted value; Step S3-6, according to the historical predicted values ​​of the number of patients visiting the target department, analyze and extract the first predicted value weight and the second predicted value weight used in calculating the predicted value of the number of patients visiting each time, and construct a weight set; Step S3-7, calculate the prediction error variance E according to the weight set, and dynamically allocate the first prediction value weight and the second prediction value weight through the smoothing factor a, and obtain the predicted value of the number of patients in the season corresponding to the target season in the current year through weighted data fusion.

6. The method for recommending insurance products based on big data according to claim 5, characterized in that: In step S4, the number of patients visiting the target department in the same quarter in different years is input into the time series prediction model to predict the number of patients in each quarter, and the total number of patients in the target department in the current year is obtained by adding the current cumulative number of patients and the predicted number of patients; the total number of patients in each target department in the current year is sorted, and the target department with the largest number of patients is selected as the recommended department.

7. The method for recommending insurance products based on big data according to claim 6, characterized in that: The specific steps of step S5 are as follows: Step S5-1, obtaining the age of patients in the recommended departments through the hospital information system, constructing a horizontal axis based on the age data, and constructing an age distribution chart based on the number of patients corresponding to different ages as the vertical axis; Step S5-2: for the constructed age distribution graph, take the set age range as the unit, count the number of patients in each age range, calculate the proportion of the number of patients in each age range to the total number of patients, and screen out the age range with the highest proportion by comparing the proportion of each age range. The people in this age range will be recommended as the objects for insurance product recommendation.

8. A big data-based insurance product recommendation system, which is applied to the big data-based insurance product recommendation method according to any one of claims 1 to 7, characterized in that: The insurance product recommendation system includes a department and time division module, a key data extraction module, a patient number prediction module, a recommended department screening module and an insurance object determination module; The department and time division module is used to subdivide the hospital departments into single departments and complete the division of each year and quarter; the key data extraction module is used to extract the target year and target quarter data from the department's visit fluctuation data; the number of patients predicted module is used to predict the number of patients in the target quarter based on historical and current data; the recommended department screening module is used to screen out the department with the largest number of patients based on the number of patients predicted results; the protection object determination module is used to determine the recommended object of the protection product through the age data of the recommended department; The output end of the department and time division module is electrically connected to the input end of the key data extraction module; the output end of the key data extraction module is electrically connected to the input end of the number of patients predicted module; the output end of the number of patients predicted module is electrically connected to the input end of the recommended department screening module; the output end of the recommended department screening module is electrically connected to the input end of the security object determination module.

9. The insurance product recommendation system based on big data according to claim 8, characterized in that: The department and time division module includes a department division unit and a quarter division unit; the department division unit is responsible for dividing all departments of the hospital into a single department and selecting a target department; the quarter division unit is used to divide each year into quarters using an equal-length time period definition method; The key data extraction module includes a target year extraction unit and a target quarter extraction unit; The target year extraction unit is used to find the year with the largest number of patients from the department visit fluctuation analysis set; The target quarter extraction unit is used to determine the quarter with the largest number of patients in the target year data; The patient number prediction module includes a basic data statistics unit and a prediction value calculation unit; the basic data statistics unit is used to obtain the corresponding quarterly data of the current year and calculate various average values ​​and cumulative number of patients; The prediction value calculation unit is used to obtain a first prediction value and a second patient number prediction value by different methods and merge them.

10. The insurance product recommendation system based on big data according to claim 8, characterized in that: The recommended department screening module includes a patient number forecasting and summarizing unit and a department screening unit; the patient number forecasting and summarizing unit is used to use a time series model to predict and summarize the number of patients in each quarter; the department screening unit is used to sort the number of patients in each department and select the recommended department; The protection object determination module includes an age distribution mapping unit and a protection object screening unit; The age distribution drawing unit is used to obtain the age data of the recommended departments to construct a distribution map; the protection object screening unit is used to divide the age ranges and screen the people in the range with the highest proportion as the recommended objects for the protection product recommendation.

Images