Supply and demand prediction method for infant nursing and supporting service facility

Abstract

The invention discloses a supply and demand prediction method for an infant nursing and supporting service facility, and the method comprises the steps: collecting the data of a newborn and a pregnant and lying-in woman in a real continuous time period, and constructing a training data set with a real infant birth rate as a label; constructing an infant birth rate prediction model, taking the training data set as input, taking the predicted infant birth rate as output, and obtaining a trained infant birth rate prediction model; according to the infant birth rate prediction model, the infant birth rate in a preset time period is pre-judged, and the demanded quantity of the rearing institutions, the matching degree of the regional facility and demand relation and the rearing utilization rate of the rearing institutions are analyzed in combination with the regional rearing characteristics. According to the method, the accuracy of infant birth rate prediction can be improved, and the growth amount of the supporting institutions, the supporting position utilization rate and the growth amount of employee teams in the future can be estimated more accurately.

Description

Technical Field

[0001] The present invention relates to the field of intelligent auxiliary decision-making technology, and in particular to a method for predicting supply and demand of infant and child care service facilities. Background Art

[0002] With the strong support of national and local infant care policies, infant care services in various places are booming. Compared with traditional industries, infant care services are an emerging service industry. New business scenario demands have emerged in the field of infant care service industry governance - intelligent governance of supply and demand of infant care service facilities.

[0003] The general process of existing infant birth rate prediction is to load data (including preprocessing, cleaning data, etc.), extract data features and model, train data, and predict using conventional statistical chi-square verification algorithms (such as Pi Dahu. Regression analysis of natural population growth rate and prediction of population birth rate based on ARIMA model [J]. Digital User, 2019. And Liu Xiaodong, Jiang Baofa. Application of residual autoregression model in China's population birth rate prediction [J]. Chinese Health Statistics, 2011, 28(3):3. DOI:10.3969 / j.issn.1002-3674.2011.03.032.). The existing infant prediction method is essentially a statistical application. The method has obvious defects. It only predicts based on pregnant women in the current year, without considering the influence of other factors, such as the number of newly married people each year and the government's birth subsidy policy, which have a great impact on the willingness to give birth.

[0004] Existing predictions of regional childcare facility expansion, childcare place utilization, or workforce development also use more traditional statistical probability analysis methods, which use statistics on past childcare facility growth patterns, fluctuations in childcare place utilization, and the average growth rate of the workforce to predict future growth trends. The results are biased and do not truly reflect the dynamic changes in business indicators.

[0005] There is currently no relevant forecasting method for the demand for this new business scenario. Summary of the invention

[0006] The purpose of the present invention is to provide a method for predicting the supply and demand of infant care and childcare service facilities, which can improve the accuracy of infant birth rate prediction and more accurately estimate the future growth of childcare institutions, the utilization rate of childcare places and the growth of the workforce.

[0007] The present invention provides the following technical solutions:

[0008] A method for forecasting the supply and demand of infant and child care service facilities, the method comprising:

[0009] (1) Collect real neonatal and maternal data in continuous time periods, use the real infant birth rate as a label, and construct a training dataset;

[0010] (2) constructing an infant birth rate prediction model, taking the training data set in step (1) as input and the predicted infant birth rate as output, and performing training to obtain a trained infant birth rate prediction model;

[0011] (3) predicting the infant birth rate within a predetermined time period based on the infant birth rate prediction model trained in step (2), and analyzing the demand for childcare institutions, the matching degree between regional facilities and demand, and the utilization rate of childcare places in childcare institutions based on the infant birth rate and regional childcare characteristics;

[0012] Among them, regional childcare characteristics include the factor G1, the ratio of the number of infants and young children to childcare institutions, the factor G2, the number of childcare workers required per infant and young child, and the factor G3, the number of childcare places required per infant and young child.

[0013] In step (1), the collected data are subjected to data fusion and data cleaning, wherein the data cleaning includes: cleaning missing values, removing unnecessary fields, filling missing contents, retrieving data, and verifying correlation.

[0014] In step (2), the infant birth rate prediction model constructed is:

[0015] r=w1x1+w2x2+b

[0016] Among them, r is the predicted infant birth rate, x1 is the number of infant births, x2 is the number of pregnant women, b is the deviation, w1 is the weight of infant births, and w2 is the weight of pregnant women;

[0017] The infant birth rate prediction model after training is:

[0018] r=Fw1x1+w2x2+b

[0019] F is the deviation factor function:

[0020] F=0.7516F1+0.2484F2

[0021] F1 is the marriage registration factor, and F2 is the fertility subsidy factor: F1 is the marriage registration factor, and F2 is the fertility subsidy factor: F1 is the absolute value of the difference in the logarithm of marriages in two consecutive years, and F2 is the ratio of the sum of the number of births in two consecutive years based on the government's subsidy policy for one, two, and multiple births.

[0022] For example, if the number of births in the subsidy policies of two consecutive years is the same, it means there is no fluctuation, and you can just take 1: for example, if the birth subsidies in a certain area in 2018 and 2019 are both for one, two, three or more children, then the birth subsidy factor in 2019 is F2 = (1+2+3) / (1+2+3) = 1; if the birth subsidy in 2018 is for one, two, three or more children, but the birth subsidy in 2019 is for two, three or more children, then the birth subsidy factor in 2019 is F2 = (2+3) / (1+2+3) = 5 / 6.

[0023] During the training process of the infant birth rate prediction model, the loss function used is the square error, and the error of the i-th sample is:

[0024]

[0025] Among them, y is the actual infant birth rate.

[0026] The method further includes fine-tuning the infant birth rate prediction model according to marriage registration and government birth subsidies, and the error sum of the infant birth rate prediction model is:

[0027]

[0028] Among them, n is the natural year number variable.

[0029] In step (2), the present invention can achieve a more accurate prediction of the infant birth rate by predicting the number of newly married people each year and the government's policy on the issuance of birth subsidies as influencing factors.

[0030] The infant birth rate prediction model is also called the regional newborn infant development trend model.

[0031] In step (3), the method for analyzing the demand for childcare institutions based on the infant birth rate and regional childcare characteristics is as follows:

[0032] (3-1-1) Demand for regional infant care institutions in the next m years:

[0033] Q m =Y m *r m *G m

[0034] Q m : The demand for childcare institutions in the next m years;

[0035] Y m : The number of pregnant women in the next m years;

[0036] r m : The predicted infant birth rate in the next m years;

[0037] G m : The ratio factor between the number of infants and young children and childcare institutions in the next m years;

[0038] (3-1-2) The average annual increase in the number of childcare institutions in the next three years:

[0039]

[0040] Y m Data estimation: According to the statistics of the Ministry of Civil Affairs released by the Industry Data Center, the number of marriage registrations has a statistical probability of deviation of about -10% in the past three years. If the number of marriage registrations this year is T, then Y m =T*0.9 m .

[0041] For example, the demand for regional infant care institutions in the next three years:

[0042] Q=Y1*r1*G+Y2*r2*G+Y3*r3*G

[0043] The average annual increase in the number of childcare institutions:

[0044]

[0045] Through the above steps, the present invention predicts the number of childcare institutions that need to be expanded in a certain area within a predetermined time period, and according to the regional newborn infant development trend model, the number of pregnant women and infants and the childcare institution matching factor, etc., it can reasonably plan the reasonable matching degree of the number of childcare institutions to be built in this area in the future period.

[0046] In step (3), the method for analyzing the matching degree between the regional facilities and demand of childcare institutions based on the infant birth rate and regional childcare characteristics is as follows:

[0047] (3-2-1) Matching degree of infant care facilities:

[0048]

[0049] P1: matching degree of infant care facilities;

[0050] Y: number of pregnant women;

[0051] r: predicted infant birth rate;

[0052] G1: The ratio factor between the number of infants and young children and childcare institutions;

[0053] M: current number of infant care facilities;

[0054] (3-2-2) Current utilization rate of infant care workers:

[0055]

[0056] P2: utilization rate of infant care practitioners;

[0057] Y: number of pregnant women;

[0058] r: predicted infant birth rate;

[0059] G2: Factor for the number of infant care workers required per capita;

[0060] M: Current number of infant care facilities.

[0061] Through the above method, the present invention can accurately predict the matching degree of the development trend of the scale of the regional workforce team building in the future period based on the existing stock of regional childcare institution practitioners, combined with the regional newborn infant development trend model and the ratio range of childcare institution practitioners to infants and young children.

[0062] In step (3), the method for analyzing the utilization rate of childcare places in childcare institutions based on the infant birth rate and regional childcare characteristics is as follows:

[0063]

[0064] S: utilization rate of support space;

[0065] Y: number of pregnant women;

[0066] r: predicted infant birth rate;

[0067] G3: Factor of the number of childcare places per capita required for infants and young children;

[0068] N: The number of support spaces in the current area.

[0069] Through the above method, the present invention can accurately predict the development trend of the utilization rate of childcare places in childcare institutions in the region in the future based on the existing stock of childcare institutions and childcare places in the region, combined with the regional newborn infant development trend model.

[0070] In step (3), the ratio factor G1 of the number of infants and young children to childcare institutions, the factor G2 of the number of infant care workers required per capita, and the factor G3 of the number of childcare places required per capita are dynamically adjusted according to the childcare subsidy funds invested by different local governments. G1, G2, and G3 are mainly realized based on the amount of childcare subsidy funds and the ratio of the base number of childcare institutions, childcare places, and personnel in the region; childcare subsidy funds are apportioned to each childcare institution, childcare place, and individual personnel, which has an incremental effect of 10% on promoting business growth.

[0071] The present invention also provides a device for predicting supply and demand of infant care and childcare service facilities, comprising a computer memory, a computer processor, and a computer program stored in the computer memory and executable on the computer processor, wherein the computer memory stores:

[0072] The preprocessing module collects real newborn and pregnant women data in continuous time periods, uses the real infant birth rate as a label, and constructs a training data set;

[0073] The infant birth rate prediction module constructs an infant birth rate prediction model, takes the training data set of the preprocessing module as input, and predicts the infant birth rate as output, and performs training to obtain the trained infant birth rate prediction model;

[0074] The childcare service facility supply and demand forecasting module predicts the infant birth rate within a predetermined time period based on the infant birth rate forecasting module, and analyzes the demand for childcare institutions, the matching degree between regional facilities and demand, and the utilization rate of childcare places in childcare institutions based on the infant birth rate and regional childcare characteristics;

[0075] Among them, regional childcare characteristics include the ratio factor G1 of the number of infants and young children to childcare institutions, the factor G2 of the number of childcare workers required per infant and young child, and the factor G3 of the number of childcare places required per infant and young child.

[0076] The infant birth rate prediction model constructed in the infant birth rate prediction module is:

[0077] r=w1x1+w2x2+b

[0078] Among them, r is the predicted infant birth rate, x1 is the number of infant births, x2 is the number of pregnant women, b is the deviation, w1 is the weight of infant births, and w2 is the weight of pregnant women;

[0079] During the training process, the infant birth rate prediction model was fine-tuned according to marriage registration and government fertility subsidies. The infant birth rate prediction model after training is:

[0080] r=Fw1x1+w2x2+b

[0081] F is the deviation factor function:

[0082] F=0.7516F1+0.2484F2

[0083] F1 is the marriage registration factor, and F2 is the birth subsidy factor: F1 is the absolute value of the difference in the logarithm of marriages in two consecutive years, and F2 is the ratio of the sum of the number of births in two consecutive years based on the government's subsidy policy for one, two, and multiple births.

[0084] The infant birth rate prediction module also includes fine-tuning the infant birth rate prediction model according to marriage registration and government birth subsidies. The error sum of the infant birth rate prediction model is:

[0085]

[0086] Among them, n is the natural year number variable.

[0087] The method for analyzing the demand for childcare institutions, the matching degree between regional facilities and demand, and the utilization rate of childcare places in the childcare service facility supply and demand forecasting module according to the infant birth rate and regional childcare characteristics is as follows:

[0088] (3-1) Demand for regional infant care institutions in the next m years:

[0089] Q m =Y m *r m *G m

[0090] Q m : The demand for childcare institutions in the next m years;

[0091] Y m : The number of pregnant women in the next m years;

[0092] r m : The predicted infant birth rate in the next m years;

[0093] G m : The ratio factor between the number of infants and young children and childcare institutions in the next m years;

[0094] (3-2) Infant care facilities matching degree P1:

[0095]

[0096] Y: number of pregnant women;

[0097] r: predicted infant birth rate;

[0098] G1: The ratio factor between the number of infants and young children and childcare institutions;

[0099] M: current number of infant care facilities;

[0100] (3-3) Utilization rate of infant care workers P2:

[0101]

[0102] (3-4) The method for the utilization rate of childcare places in childcare institutions is S:

[0103]

[0104] N: The number of support spaces in the current area.

[0105] The Y m According to the statistical bulletin of the Ministry of Civil Affairs on the development of civil affairs released by the Industry Data Center, the deviation of the number of marriage registrations in the past three years is about -10%. Calculation: Let the number of marriage registrations in the next m years be T, then Y m =T*0.9 m The factor G1 for the ratio of the number of infants and young children to childcare institutions, the factor G2 for the number of childcare workers required per infant and young child, and the factor G3 for the number of childcare places required per infant and young child are dynamically adjusted based on the childcare subsidy funds invested by different local governments.

[0106] The present invention has the following advantages:

[0107] The prediction method provided by the present invention can achieve a more accurate prediction of the infant birth rate, and then based on the predicted infant birth rate and combined with regional childcare characteristics, more accurately estimate the future growth of childcare institutions, the utilization rate of childcare places, and the growth of the workforce.

[0108] The prediction method provided by the present invention provides scientific, efficient, accurate and intelligent technical support for the business scenarios of supply and demand management of infant service facilities in the infant care service industry, supports and improves the supply and demand prediction capability of childcare service facilities, is conducive to the realization of intelligent supply and demand management of infant care and childcare service facilities, and provides a basis for the supply and demand balance of childcare service facilities for relevant government departments to implement decisions, such as providing a decision-making basis for the construction of infant care service facilities and the training of childcare practitioners by relevant government health departments. BRIEF DESCRIPTION OF THE DRAWINGS

[0109] Figure 1 A flow chart of a method for predicting supply and demand of infant and child care service facilities provided by the present invention;

[0110] Figure 2 Create a logical flow chart for data cleaning. DETAILED DESCRIPTION

[0111] The specific technical scheme of the present invention will be further described below in conjunction with the accompanying drawings.

[0112] like Figure 1 As shown, the method for predicting the supply and demand of infant care and childcare service facilities provided by the present invention comprises the following steps:

[0113] S1. Collect real neonatal and maternal data in continuous time periods, use the real infant birth rate as labels, and construct a training dataset.

[0114] Before building a prediction model for infant birth rate, it is necessary to extract and clean government public resource population data to obtain standardized basic data. The data model and algorithm for infant birth rate are based on the establishment of a basic infant database. The specific process includes:

[0115] S1-1. Data source integration

[0116] Data collection is done on government-related data resources, using database data synchronization. The data synchronization time strategy rules can be customized (daily, weekly, monthly, quarterly, and annual). The data synchronization method can be set to synchronize full coverage of business data, incremental synchronization, data correction, and other synchronization methods as needed.

[0117] The data source configuration uses the integrated data source framework of the microservice framework SpringBoot to configure a bidirectional source database. The configuration parameters mainly include: database server IP address, port, account, password, source table, target table, synchronization field, etc.

[0118] The data source fusion module structure mainly includes the following components:

[0119] ■JSON format data source construction

[0120] This component is adapted to encapsulate RDBMS (relational database) data sources, MongoDB non-relational data, and supports the construction of big data analysis database format data in JSON format. The construction of data source JSON format adopts the extensible design method of factory mode, which can flexibly cope with the data formats of different types of data sources.

[0121] ■Task creation component

[0122] The task creation component can create various strategy mode data source fusions, including full, incremental, and change. It supports the creation of custom time tasks, mainly using time regular expressions to efficiently express complex time strategies and frequency strategies. Flexible time strategies can adapt to data fusion tasks required by various application environments.

[0123] ■Log management component

[0124] According to different log types, such as data source fusion task scheduling log, statistical data fusion record log, data fusion process log, request queue log, etc., log management can be used to query the detailed information of data fusion process execution tasks in real time and efficiently, and can also assist in data fusion troubleshooting so that the process can be "mirrored and restored".

[0125] ■Data fusion task scheduling queue component

[0126] High-concurrency data fusion tasks are mainly triggered by the task management component according to the task time strategy. The task manager puts the relevant parameter information of the data fusion task (data source, original table, target table, fusion mode) into the data fusion task scheduling queue. The task scheduling queue is responsible for monitoring the queue capacity, parsing the data fusion task parameters, and calling the data fusion task module. The task scheduling queue component is a conventional solution to deal with high concurrency in the system.

[0127] ■Task execution management component

[0128] The task execution management component receives the task request from the task scheduling queue, parses the request parameters, and delegates the specific task object to perform the data fusion task. The task execution object is automatically executed by the task creation component during system initialization and is generated according to the various task configuration files.

[0129] S1-2. Data cleaning

[0130] The present invention specifically formulates special data cleaning business components based on business characteristics. The data business characteristics extract information including: childcare institution registration information, childcare institution practitioners, childcare institution practitioners' qualification certificates, childcare institution evaluations, infant and young children files, pregnant women files, cities, districts, counties, towns and streets, etc.; integrity and consistency checks are performed on various types of business data; legality checks are performed on data field types (whether it is empty, integer, floating point, text type), and data types are standardized and transformed and adapted. The field type of the corresponding database can be matched according to different types of data sources. The data cleaning process is as follows:

[0131] S1-2-1. Clean up missing values

[0132] The first step in data cleaning is to clean missing values. Missing values ​​are a very common data problem, and there are many ways to deal with them. Here is a very common method. First, clarify the scope of missing values: calculate the missing value ratio for each field, and formulate strategies based on the missing ratio and field importance.

[0133] S1-2-2. Remove unnecessary fields

[0134] The highlight of this process is the trial and error method, which is to back up the data first, then conduct a small-scale data experiment, and then apply it to a large amount of data after confirming that it is correct. This method can avoid data loss caused by the inability to restore deleted data.

[0135] S1-2-3. Fill in missing content

[0136] There are three ways to fill missing data: filling with business knowledge / experience, filling with the results calculated with the same indicator, and filling with the results calculated with different indicators.

[0137] S1-2-4, Re-acquisition

[0138] Re-collecting data is for data with important indicators but a high missing rate. This requires consulting the data collector or business personnel, or obtaining relevant data from other channels.

[0139] S1-2-5. Correlation Verification

[0140] If there are multiple sources of data, it is necessary to verify the correlation.

[0141] The data cleaning of the present invention has unique highlights: it supports multi-threading, can process massive data without pressure, and can effectively improve the performance of data processing. The powerful data processing function not only supports heterogeneous data, but also has built-in data preprocessing functions such as sorting, deduplication, mapping, row and column merging, row and column conversion aggregation, and null value removal.

[0142] The data cleaning logic flow chart is as follows: Figure 2 As shown, data cleaning rule configuration includes: null value specification, type conversion, data consistency, length verification, and field mapping matching.

[0143] S2. Construct an infant birth rate prediction model, take the training data set in step (1) as input, and the predicted infant birth rate as output, and perform training to obtain a trained infant birth rate prediction model.

[0144] The specific process is:

[0145] Through the birth data of newborn infants and young children in the past five years and the data of pregnant women, the implicit information in the data is obtained through the processes of collection, storage, cleaning, calculation, etc., and further through data analysis, a data analysis model is established to verify, screen, repeatedly calculate, judge and other operations on the data, and compare and analyze the ideal situation such as target data with the actual situation to find the internal rules. The present invention selects the infant birth variable and the pregnant woman variable element to predict and estimate the relationship between the future birth rate and these two elements.

[0146] According to the linear programming algorithm, the following model is established:

[0147] r=w1x1+w2x2+b

[0148] The model training and prediction model and related algorithms for infant birth are demonstrated as follows:

[0149] S2-1. Model training and algorithm

[0150] An important element of model training is the training data set. This embodiment uses a series of historical real data as the training data set. Each piece of data in the training data set is used as a sample, and the result in the data set is used as a label. Each element feature in the data set. For example, this module collects real newborn and pregnant women data sets in a continuous time period, in which each newborn and pregnant woman and maternal information is one of our samples. Birth rate and waiting rate are data element features.

[0151] Assume that the number of samples collected is n, and the characteristics of the i-th sample are Tagged with

[0152] The birth probability of the i-th infant or pregnant woman is:

[0153]

[0154] Among them, r is the predicted infant birth rate, x1 is the number of infant births, x2 is the number of pregnant women, b is the deviation, w1 is the weight of infant births, and w2 is the weight of pregnant women.

[0155] S2-2. Model prediction and algorithm

[0156] After the model training is completed, a set of This solution is not necessarily the optimal solution with the minimum error, but it is an approximation of the optimal solution. In this way, the trained model can be used to predict the supply scale of regional childcare service facilities. We call this model a prediction model:

[0157]

[0158] The purpose of model training is to obtain a set of It can make the predicted value closer to the accurate value; and use a function as the error between the accurate value and the predicted value, which we call the loss function. The smaller the value calculated by the loss function, the closer the predicted value is to the true value, and thus the better the training model. This embodiment uses the square error as the loss function. The error of the i-th sample in this example can be:

[0159]

[0160] When taking derivatives, the square can cancel each other out, thus simplifying the calculation. Therefore, the sum of all errors in the sample is used as a basis for the quality of the training model.

[0161] In the present invention, the error sum of the model is:

[0162]

[0163] The present invention obtains the weight of infant births, the weight of pregnant women, and the deviation through historical data training, and makes appropriate fine adjustments according to the deviations of annual marriage registrations in the region, government birth subsidies, etc.

[0164] The deviation factor function is as follows:

[0165] F=0.7516F1+0.2484F2, F1 is the marriage registration factor, and F2 is the birth subsidy factor.

[0166] The present invention defines the external influencing factors as the deviation factor function, and the specific implementation process is as follows:

[0167] selector=ft.ChiSqSelector(4).fit(births_train)

[0168] topFeatures_train=(births_train.map(lambda row:row.label)\

[0169] .zip(selector.transform(births_train.map(lambda row:row.features)))

[0170] ).map(lambda row:reg.LabeledPoint(row[0],row[1]))

[0171] This code uses the algorithm model implemented using the MLlib classification feature library of the open source spark framework.

[0172] The infant prediction model and algorithm provided by the present invention have a significantly higher accuracy in predicting infant birth rates in Zhejiang Province in previous years than other existing predictions.

[0173] Combining the data in Table 1 and Table 2, taking Zhejiang Province in 2020 as an example, the birth rate of the whole province is 7.13‰, which is 3.14 percentage points lower than the birth rate of 10.27‰ in 2010, and the decline is quite significant. The existing conventional algorithm calculates the birth rate of 2020 to be 10.0‰, while the infant birth rate prediction model algorithm of the present invention calculates the birth rate of 2020 to be 7.38‰, which is very close to the actual birth rate of this year. Specifically:

[0174] The traditional algorithm uses the growth rate deviation of the previous year 2019 and the year before 2018 to estimate the growth rate deviation of 2021 and 2020, Δ 2018-2019 =-0.51, from which we can conclude that the population birth rate in 2020 = 10.51-0.51 = 10.0‰.

[0175] In this embodiment, the sum of the number of births for the subsidy policy for one, two, and multiple births in 2019 and 2020 is the same, so F2 is 1; then according to the logarithm of marriage registrations in Zhejiang Province in Table 2, according to the prediction method provided by the present invention, the F deviation factor (F=0.7516F1+0.2484F2) is first calculated from the number of marriage registrations, and the deviation factor F in 2020 is F=0.7516*(29.36-27.28)+0.2484*(1)≈1.812.

[0176] According to the formula n is taken as 1, w1 infant birth weight factor is taken as 0.75, w2 maternal weight factor is taken as 0.65, y predicted infant birth rate deviation factor is taken as 0.5, and the calculated error sum is b=-2.8.

[0177] According to the formula i is 1, and the above formula F≈=1.812, b=-2.8, w1=0.75, w2=0.65, x1 infant births in 2019 is about 29.36, x2 maternal births is about 29.36, and the 2020 population birth rate is calculated to be 7.38‰. The actual 2020 population birth rate is 7.13‰.

[0178] Based on the above reasoning, it can be concluded that the estimated infant birth rate of the present invention is more accurate than the traditional method.

[0179] Table 1 Natural changes in population from Zhejiang Province Statistical Yearbook

[0180] years Resident population at the end of the year Births Birth rate (‰) 2014 5890 57.8 10.51 2015 5985 58.1 10.52 2016 6072 62.4 11.22 2017 6170 67 11.92 2018 6273 62.8 11.02 2019 6375 60.9 10.51 2020 6468 45.8 7.13 2021 6540 44.9 6.9

[0181] Table 2 Number of marriage registrations in Zhejiang

[0182] years Number of marriages 2014 436,800 2015 392,500 2016 366,800 2017 344,400 2018 337,300 2019 293,600 2020 272,800 2021 304,300

[0183] S3. The trained infant birth rate prediction model predicts the infant birth rate within a predetermined time period, and analyzes the demand for childcare institutions, the matching degree between regional facilities and demand, and the utilization rate of childcare places in childcare institutions based on the infant birth rate and regional childcare characteristics; among which, the regional childcare characteristics include the ratio factor G1 of the number of infants and childcare institutions, the factor G2 of the number of infant care practitioners required per infant, and the factor G3 of the number of childcare places required per infant.

[0184] S3-1. Analyze the demand for childcare institutions based on infant birth rate and regional childcare characteristics

[0185] S3-1-1. Demand for regional infant care institutions in the next m years:

[0186] Q m =Y m *rm *G m

[0187] Q m : The demand for childcare institutions in the next m years;

[0188] Y m : The number of pregnant women in the next m years;

[0189] r m : The predicted infant birth rate in the next m years;

[0190] G m : The ratio factor between the number of infants and young children and childcare institutions in the next m years.

[0191] S3-1-2. The average annual increase in the number of childcare institutions in the next three years:

[0192]

[0193] Y m Data estimation: According to the statistics of the Ministry of Civil Affairs released by the Industry Data Center, the number of marriage registrations has a statistical probability of deviation of about -10% in the past three years. If the number of marriage registrations this year is T, then Y m =T*0.9 m .

[0194] Taking Hangzhou as an example, according to the traditional algorithm based on the data in 2022, it is predicted that the scale of childcare institutions in 2023 will be 1,163. The present invention predicts that the demand for childcare institutions will be 1,359. The actual number of childcare institutions in Hangzhou in 2023 will be 1,311. Therefore, the number of childcare institutions predicted by the present invention is more accurate.

[0195] According to the formula Q=Y1*r1*G+Y2*r2*G+Y3*r3*G for predicting the demand for childcare institutions in the next three years, the ratio factor G1 of the number of infants and young children to childcare institutions is taken as 12, and Y1, r1, Y2, and r2 are based on the principles of the above reasoning process. It is easy to calculate that the annual demand for childcare institutions in the next three years will be 4077, and the average annual demand is 4077 / 3=1359.

[0196] The traditional algorithm calculates that the number of childcare institutions in 2023 will be 1,163, based on 1,058 in 2022 and an incremental increase of 10%.

[0197] By comparing the present invention with the traditional simple algorithm, it is found that the present invention is more accurate in predicting the demand for childcare institutions in the next m years.

[0198] S3-2. Analyze the matching degree between the demand for childcare institutions and regional facilities based on the infant birth rate and regional childcare characteristics

[0199] S3-2-1. Matching degree of infant care facilities:

[0200]

[0201] P1: matching degree of infant care facilities;

[0202] Y: number of pregnant women;

[0203] r: predicted infant birth rate;

[0204] G1: The ratio factor between the number of infants and young children and childcare institutions;

[0205] M: current number of infant care facilities;

[0206] S3-2-2. Utilization rate of infant care workers:

[0207]

[0208] P2: utilization rate of infant care practitioners;

[0209] Y: number of pregnant women;

[0210] r: predicted infant birth rate;

[0211] G2: Factor for the number of infant care workers required per capita;

[0212] M: Current number of infant care facilities.

[0213] S3-3. Analysis of the utilization rate of childcare places in childcare institutions based on infant birth rate and regional childcare characteristics

[0214]

[0215] S: utilization rate of support space;

[0216] Y: number of pregnant women;

[0217] r: predicted infant birth rate;

[0218] G3: Factor of the number of childcare places per capita required for infants and young children;

[0219] N: The number of support spaces in the current area.

[0220] The present invention also provides a software architecture for implementing the above-mentioned method for predicting supply and demand of infant care and childcare service facilities, including microservices, data synchronization, log collection, distributed task scheduling, message queues, etc.

[0221] a. Microservices

[0222] Microservices provide a basic framework for distributed system services, monitor service clusters, implement service registration management, and automate service discovery, and can handle service failures without affecting business functions. For example, if there are three microservices for the annual demand for services in infant care facilities in the next three years, and microservices support dynamic expansion, then the annual demand for services in infant care facilities in the next five years can be implemented online without stopping the application.

[0223] b. Data synchronization

[0224] Data synchronization is the foundation of the data fusion business of this system. It is a customizable and scalable data cleaning function based on business needs, including data standard formulation, data conversion, and data format verification, and finally pushes the data that meets the requirements to the target database. In conjunction with the data caching service, data synchronization performance can be improved. In the process of extracting government population public data resources and integrating cost applications into the infant basic database, there are inconsistencies in data formats between different systems and inconsistent database systems. Data synchronization needs to solve this problem.

[0225] c. Log management

[0226] The microservices of this system establish a set of unified log standards, log formats, log storage strategies, log subject objects, etc., and implement log monitoring strategies based on the log collection framework, which can warn of log anomalies in real time, assist in quickly locating business logic links for system failures, quickly analyze the causes of problems, and enhance system robustness. During the operation of this application, modules such as data synchronization, infant birth rate prediction, and matching of infant care service facilities and personnel need to generate key and necessary operation logs at all times. The significance of logs is to facilitate the investigation of related problems and the details of the monitoring module operation process.

[0227] d. Distributed task scheduling

[0228] Distributed task scheduling uses multiple servers to execute scheduling tasks, increase task concurrency, and improve task execution efficiency. The task scheduling center manages each task execution plan, manages the task execution order, execution result notification, and execution log archiving. An important function of distributed task scheduling is to handle task failure strategies, retry strategies, execution compensation strategies, task emergency termination strategies, etc. Task execution is summarized in a unified manner, and the task execution status can be viewed at a glance. During the operation of this application, there will be a series of scheduled tasks. During the data synchronization process, according to the actual needs of the government population resource system, it may not be executed until the evening. At the same time, according to the operating pressure of the system, the execution data intensity of the task also has different requirements. Distributed task scheduling can meet flexible and diverse task management.

[0229] e.Message Queue

[0230] It is a mechanism for transmitting, storing and processing messages in software systems. It is often used for asynchronous communication, allowing different components or systems to be decoupled in time so that they can produce and consume messages independently. The basic principle of a message queue is to send messages from one sender to one or more receivers. The sender puts the message into the queue, and the receiver gets the message from the queue for processing. In the b. data synchronization process, the number of infants and pregnant women in the past 10 years needs to be executed concurrently. The data synchronization for each year can be set to a thread for processing, so 10 messages are required to trigger each task.

[0231] Corresponding to the embodiment of the method for predicting the supply and demand of infant care and childcare service facilities described above, the present invention further provides an apparatus for predicting the supply and demand of infant care and childcare service facilities, comprising a computer memory, a computer processor, and a computer program stored in the computer memory and executable on the computer processor, wherein the computer memory stores:

[0232] The preprocessing module collects real newborn and pregnant women data in continuous time periods, uses the real infant birth rate as a label, and constructs a training data set;

[0233] The infant birth rate prediction module constructs an infant birth rate prediction model, takes the training data set of the preprocessing module as input, and predicts the infant birth rate as output, and performs training to obtain the trained infant birth rate prediction model;

[0234] The childcare service facility supply and demand forecasting module predicts the infant birth rate within a predetermined time period based on the infant birth rate forecasting module, and analyzes the demand for childcare institutions, the matching degree between regional facilities and demand, and the utilization rate of childcare places in childcare institutions based on the infant birth rate and regional childcare characteristics;

[0235] Among them, regional childcare characteristics include the ratio factor G1 of the number of infants and young children to childcare institutions, the factor G2 of the number of childcare workers required per infant and young child, and the factor G3 of the number of childcare places required per infant and young child.

[0236] The embodiment of the present invention in combination with the method for predicting the supply and demand of infant care service facilities can be applied to any device with data processing capabilities, which can be a device or apparatus such as a computer. The device embodiment can be implemented by software, or by hardware or a combination of software and hardware.

[0237] The implementation process of the functions and effects of each unit in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here.

[0238] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The device embodiment described above is only schematic, in which the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the scheme of the present invention. Ordinary technicians in this field can understand and implement it without paying creative work.

[0239] An embodiment of the present invention also provides a computer-readable storage medium having a program stored thereon. When the program is executed by a processor, the method for predicting supply and demand of infant and child care service facilities in the above embodiment is implemented.

[0240] The computer-readable storage medium may be an internal storage unit of any device with data processing capability described in any of the aforementioned embodiments, such as a hard disk or a memory. The computer-readable storage medium may also be an external storage device of any device with data processing capability, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), an SD card, a flash card (Flash Card), etc. equipped on the device. Furthermore, the computer-readable storage medium may also include both an internal storage unit and an external storage device of any device with data processing capability. The computer-readable storage medium is used to store computer programs and other programs and data required by any device with data processing capability, and may also be used to temporarily store data that has been output or is to be output.

[0241] Those skilled in the art can understand that the above is only a single example of the invention and is not intended to limit the invention. Although the invention is described in detail with reference to the above examples, those skilled in the art can still modify the technical solutions recorded in the above examples or replace some of the technical features therein with equivalents. Any modification, equivalent replacement, etc. made within the spirit and principle of the invention shall be included in the protection scope of the invention.

Claims

1. A method for predicting supply and demand of infant care and childcare service facilities, characterized in that: The method comprises: (1) Collect real neonatal and maternal data in continuous time periods, use the real infant birth rate as a label, and construct a training dataset; (2) constructing an infant birth rate prediction model, taking the training data set in step (1) as input and the predicted infant birth rate as output, and performing training to obtain a trained infant birth rate prediction model; (3) predicting the infant birth rate within a predetermined time period based on the infant birth rate prediction model trained in step (2), and analyzing the demand for childcare institutions, the matching degree between regional facilities and demand, and the utilization rate of childcare places in childcare institutions based on the infant birth rate and regional childcare characteristics; Among them, regional childcare characteristics include the ratio factor G1 of the number of infants and young children to childcare institutions, the factor G2 of the number of childcare workers required per infant and young child, and the factor G3 of the number of childcare places required per infant and young child.

2. The method for predicting supply and demand of infant care and childcare service facilities according to claim 1, characterized in that: In step (1), the collected data are subjected to data fusion and data cleaning, wherein the data cleaning includes: cleaning missing values, removing unnecessary fields, filling missing contents, retrieving data, and verifying correlation.

3. The method for predicting supply and demand of infant care and childcare service facilities according to claim 1, characterized in that: In step (2), the infant birth rate prediction model constructed is: r=w1x1+w2x2+b Among them, r is the predicted infant birth rate, x1 is the number of infant births, x2 is the number of pregnant women, b is the deviation, w1 is the weight of infant births, and w2 is the weight of pregnant women; During the training process, the infant birth rate prediction model was fine-tuned according to marriage registration and government birth subsidies. The infant birth rate prediction model after training is: r=Fw1x1+w2x2+b F is the deviation factor function: F=0.7516F1+0.2484F2 F1 is the marriage registration factor, and F2 is the birth subsidy factor: F1 is the absolute value of the difference in the logarithm of marriages in two consecutive years, and F2 is the ratio of the sum of the number of births in two consecutive years based on the government's subsidy policy for one, two, and multiple births.

4. The method for predicting supply and demand of infant care and childcare service facilities according to claim 3, characterized in that: The error sum of the infant birth rate prediction model is: Among them, n is the natural year number variable.

5. The method for predicting supply and demand of infant care and childcare service facilities according to claim 1, characterized in that: In step (3), the method for analyzing the demand for childcare institutions based on the infant birth rate and regional childcare characteristics is as follows: (3-1) Demand for regional infant care institutions in the next m years: Q m =Y m *r m *G m Q m : The demand for childcare institutions in the next m years; Y m : The number of pregnant women in the next m years; r m : The predicted infant birth rate in the next m years; G m : The ratio factor between the number of infants and young children and childcare institutions in the next m years.

6. The method for predicting supply and demand of infant care and childcare service facilities according to claim 5, characterized in that: In step (3), the method for analyzing the matching degree between regional facilities and demand based on infant birth rate and regional childcare characteristics is as follows: (3-2) Infant care facilities matching degree P1: Y: number of pregnant women; r: predicted infant birth rate; G1: The ratio factor between the number of infants and young children and childcare institutions; M: current number of infant care facilities; (3-3) Utilization rate of infant care workers P2:

7. The method for predicting supply and demand of infant care and childcare service facilities according to claim 6, characterized in that: In step (3), the method for analyzing the utilization rate of childcare places in childcare institutions based on the infant birth rate and regional childcare characteristics is as follows: (3-4) Utilization rate of childcare places in childcare institutions S: N: The number of support spaces in the current area.

8. The method for predicting supply and demand of infant care and childcare service facilities according to claim 6, characterized in that: The Y m According to the statistical bulletin of the Ministry of Civil Affairs on the development of civil affairs released by the Industry Data Center, the deviation of the number of marriage registrations in the past three years is about -10%. Calculation: Let the number of marriage registrations in the next m years be T, then Y m =T*0.9 m .

9. The method for predicting supply and demand of infant care and childcare service facilities according to claim 1, characterized in that: The factor G1 for matching the number of infants and young children to childcare institutions, the factor G2 for the number of childcare workers required per infant and young child, and the factor G3 for the number of childcare places required per infant and young child are dynamically adjusted based on the childcare subsidy funds invested by different local governments.

10. A device for predicting supply and demand of infant care and childcare service facilities using the method of any one of claims 1 to 9, comprising a computer memory, a computer processor, and a computer program stored in the computer memory and executable on the computer processor, wherein the computer memory stores: The preprocessing module collects real newborn and pregnant women data in continuous time periods, uses the real infant birth rate as a label, and constructs a training data set; The infant birth rate prediction module constructs an infant birth rate prediction model, takes the training data set of the preprocessing module as input, and predicts the infant birth rate as output, and performs training to obtain the trained infant birth rate prediction model; The childcare service facility supply and demand forecasting module predicts the infant birth rate within a predetermined time period based on the infant birth rate forecasting module, and analyzes the demand for childcare institutions, the matching degree between regional facilities and demand, and the utilization rate of childcare places in childcare institutions based on the infant birth rate and regional childcare characteristics; in, Regional childcare characteristics include the ratio factor G1 between the number of infants and young children and childcare institutions, the factor G2, the number of childcare workers required per infant and young child, and the factor G3, the number of childcare places required per infant and young child.

Images