A method and system for implementing a formula recommendation algorithm

Abstract

The present invention relates to an implementation method and system of a formula recommendation algorithm, and belongs to the technical field of formula recommendation. The method includes: pre-storing several calculation formulas for fund operations, the several calculation formulas include conventional formulas and / or custom formulas; fund operations include salary operations and / or engineering operations; obtaining the data source corresponding to each fund operation and determining the corresponding recommendation conditions; the data source corresponding to the salary operation is attendance data, and the data source corresponding to the engineering operation is engineering management data; the recommendation conditions are conditions set according to demand; the data source is brought into different calculation formulas for the corresponding fund operations, and the calculation results of different calculation formulas are compared; and the calculation formula corresponding to the calculation result that meets the recommendation conditions is recommended. The present invention does not require long-term learning by the staff, and only requires setting the recommendation conditions to obtain the optimal calculation formula, which is simple to operate and highly efficient.

Description

Technical Field

[0001] The present invention relates to a method and system for implementing a formula recommendation algorithm, and belongs to the technical field of formula recommendation. Background Art

[0002] Formula calculations are often used in the business systems of each company. Each item has several different calculation formulas. For example, the calculation formula for salary calculation involves calculation items such as personal leave hours, overtime hours, basic salary, education subsidy, absenteeism hours, monthly working days, etc. There are many different calculation formulas for different calculation items, which makes the calculation formulas for salary calculations different. Different calculation formulas get different calculation results, which makes the company's operating costs different.

[0003] It is very important for enterprise operation and management to find the best and most suitable calculation formula among these different calculation formulas. The selection of the existing optimal formula generally requires staff to learn different calculation formulas for a long time, or to design corresponding calculation formulas through learning, and to determine the optimal formula by comparing different calculation results. The existing method not only increases the learning cost of the enterprise, but also has low overall efficiency. Summary of the invention

[0004] The purpose of this application is to provide a method and system for implementing a formula recommendation algorithm to solve the problems of low efficiency and high cost in the existing formula determination process.

[0005] To achieve the above objectives, the present application proposes a technical solution for implementing a formula recommendation algorithm, comprising the following steps:

[0006] 1) Pre-store several calculation formulas for fund operations, including conventional formulas and / or custom formulas; conventional formulas are unified calculation formulas developed based on expert experience; custom formulas are calculation formulas built based on demand; fund operations include salary operations and / or engineering operations;

[0007] 2) Obtain the data source corresponding to each fund operation and determine the corresponding recommended conditions; the data source corresponding to the salary operation is the attendance data, and the data source corresponding to the engineering operation is the engineering management data; the recommended conditions are conditions set according to the needs;

[0008] 3) Bring the data source into the corresponding different calculation formulas for fund operation and compare the calculation results of different calculation formulas;

[0009] 4) Recommend the calculation formula corresponding to the calculation result that meets the recommendation conditions.

[0010] The technical solution of the implementation method of the formula recommendation algorithm of the present invention has the beneficial effect that: the present invention stores several calculation formulas corresponding to each fund operation, obtains the data source corresponding to each fund operation when recommendation is needed, brings the data source into different calculation formulas, obtains different calculation results, compares the calculation results, and selects the calculation formula that meets the recommendation conditions for recommendation. The present invention does not require long-term learning by the staff, and only needs to set the recommendation conditions to automatically obtain the optimal calculation formula, which is simple to operate and highly efficient.

[0011] Furthermore, the custom formula is constructed through a formula definition module, which includes operators and functions; the operators include arithmetic operators, comparison operators and logical operators; the functions include sine function, cosine function, tangent function, absolute value function and judgment function.

[0012] In order to avoid the complicated process of defining formulas by formula programming, formulas are built through the formula definition module to simplify the process of formula input. The simple formula building method is suitable for workers with less experience and can meet the work needs of workers at all levels.

[0013] Furthermore, the recommendation conditions are adjusted according to time factors and operational factors, and the recommendation results are updated after the recommendation conditions are adjusted.

[0014] When adjustments need to be made to capital operations at different times or due to the impact of company operations, you only need to adjust the recommended conditions to get the corresponding optimal formula, which is convenient and simple.

[0015] Furthermore, the conventional formula is updated regularly.

[0016] When a new general formula is found, the latest general formula can be stored to meet the latest needs.

[0017] In addition, the present application also proposes a technical solution for an implementation system of a formula recommendation algorithm, wherein the implementation system of the formula recommendation algorithm includes an intelligent formula learning library, a calculation comparison module and a recommendation module, wherein the intelligent formula learning library is used to store several calculation formulas for financial operations; the several calculation formulas include conventional formulas and / or custom formulas; conventional formulas are unified calculation formulas formulated based on expert experience; custom formulas are calculation formulas built by themselves according to needs; financial operations include salary operations and / or engineering operations; the calculation comparison module is used to obtain the data source corresponding to each financial operation; the data source is brought into the different calculation formulas of the corresponding financial operations, and the calculation results of different calculation formulas are compared; the data source corresponding to the salary operation is attendance data, and the data source corresponding to the engineering operation is engineering management data; the recommendation module is used to determine the recommendation conditions corresponding to each financial operation, and recommend the calculation formula corresponding to the calculation result that meets the recommendation conditions; the recommendation conditions are conditions set according to needs.

[0018] The beneficial effect of the technical solution of the implementation system of the formula recommendation algorithm of the present invention is: the present invention stores several calculation formulas corresponding to various fund operations through the intelligent formula learning library. When recommendation is needed, the calculation comparison module obtains the data source corresponding to each fund operation, brings the data source into different calculation formulas, obtains different calculation results, compares the calculation results, and the recommendation module selects the calculation formula that meets the recommendation conditions for recommendation. The present invention does not require long-term learning by the staff, and only needs to set the recommendation conditions to automatically obtain the optimal calculation formula, which is simple to operate and efficient.

[0019] Furthermore, it also includes a formula definition module, through which custom formulas are built. The formula definition module includes operators and functions; operators include arithmetic operators, comparison operators and logical operators; functions include sine function, cosine function, tangent function, absolute value function and judgment function.

[0020] In order to avoid the complicated process of defining formulas by formula programming, formulas are built through the formula definition module to simplify the process of formula input. The simple formula building method is suitable for workers with less experience and can meet the work needs of workers at all levels.

[0021] Furthermore, the recommendation conditions are adjusted according to time factors and operational factors, and the recommendation results are updated after the recommendation conditions are adjusted.

[0022] When adjustments need to be made to capital operations at different times or due to the impact of company operations, you only need to adjust the recommended conditions to get the corresponding optimal formula, which is convenient and simple.

[0023] Furthermore, the conventional formula is updated regularly.

[0024] When a new general formula is found, the latest general formula can be stored to meet the latest needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a structural block diagram of the implementation system of the formula recommendation algorithm of the present invention. DETAILED DESCRIPTION

[0026] Implementation example of the formula recommendation algorithm system:

[0027] The main idea of ​​the present invention is that, based on the problem of high cost and low efficiency of the existing optimal calculation formula selection method, the present invention stores various calculation formulas for financial operations in an intelligent formula learning library. After obtaining the corresponding data source, the calculation comparison module brings the data source into different calculation formulas to obtain different calculation results. The recommendation module screens and recommends the calculation formula according to the set recommendation conditions and calculation results, so as to recommend the optimal calculation formula more simply and efficiently.

[0028] Specifically, the implementation system of the formula recommendation algorithm is as follows Figure 1 As shown, it includes a formula definition module, an intelligent formula learning library, a calculation comparison module, and a recommendation module.

[0029] The formula definition module is used to build calculation formulas, build custom formulas for salary calculations and engineering calculations; the intelligent formula learning library is used to store the built custom formulas or obtained regular formulas, including the salary business formula library and the engineering budget business formula library; the calculation comparison module (also known as Figure 1 The calculation result comparison center in the project) includes an accounting module and a budget module. The accounting module is used to calculate and compare the calculation results of different calculation formulas corresponding to salary operations, and the budget module is used to calculate and compare the calculation results of different calculation formulas corresponding to engineering operations; the recommendation module (also known as Figure 1 The intelligent formula recommendation module in the example is used to recommend the calculation formula corresponding to the calculation result that meets the recommendation condition according to the recommendation condition and the calculation result obtained by the calculation comparison module.

[0030] The implementation system of the formula recommendation algorithm is used to implement the implementation method of the formula recommendation algorithm, and the method includes the following steps:

[0031] 1) Build several custom formulas for fund operations through the formula definition module, and obtain several conventional formulas for fund operations.

[0032] Conventional formulas are unified calculation formulas developed based on expert experience; custom formulas are calculation formulas built by the formula definition module based on demand; financial operations include salary operations and engineering operations, that is, conventional formulas include several calculation formulas for salary operations and several calculation formulas for engineering operations.

[0033] The formula definition module includes the definition of the operation object, function, and operator; through the combination of functions, operation objects, and operators, several small formulas are formed to determine the big formula. For the operation object of salary operation, the small formulas include the holiday overtime pay formula, meal allowance formula, etc. The big formula is the calculation formula of salary operation, which is the formula for payable wages. Different calculation methods of small formulas will result in several calculation formulas in the big formula.

[0034] For example, the formula for holiday overtime pay is: holiday overtime pay = basic salary / 21.75 / 8*3*holiday overtime hours; the formula for meal allowance is: meal allowance = iff(whether to pay meal allowance == "yes", number of days worked*12,0); when the iff function determines whether to pay meal allowance, it pays meal allowance at the rate of 12 yuan per day based on the actual number of days worked.

[0035] The formula for wages payable is: wages payable = basic salary + position salary + overtime pay + meal allowance.

[0036] Operators include arithmetic operators (addition, subtraction, multiplication and division), comparison operators, logical operators (> greater than, >= greater than or equal to, < less than, <= less than or equal to, == equal to, != not equal to), etc. The operator priority is consistent with the general order.

[0037] Functions are mathematical functions, including sine function math.sin(d), cosine function math.cos(d), tangent function math.tan(d), absolute value function math.abs(d), and judgment function iff(true, true, false), which judges the truth or falsehood through the first value and executes subsequent formula blocks.

[0038] The formula definition module supports common types such as numbers, Booleans, strings, integers, BgiDecimal, and more.

[0039] The list of operators is shown in Table 1:

[0040] Table 1 List of operators

[0041] Priority Operators Combination 1 ()[]. From left to right 2 ！～ Right to Left 3 * / ％ From left to right 4 +- From left to right 5 <<>>>>> From left to right 6 <<＝>>＝ From left to right 7 ＝＝！＝ From left to right 8 & From left to right 9 ^ From left to right 10 | From left to right 11 && From left to right 12 || From left to right 13 ？: From left to right 14 ＝＝～ Right to Left

[0042] The list of functions is shown in Table 2:

[0043] Table 2 Function list

[0044]

[0045]

[0046] 2) Store various custom formulas and conventional formulas for fund operations in the intelligent formula learning library.

[0047] from Figure 1 It can be seen that the intelligent formula learning library includes the salary business formula library and the engineering budget business formula library, and there may also be other expanded business formula libraries.

[0048] The salary business formula library stores the big and small formulas for salary calculations, and different positions have different payable wages. The engineering budget business formula library stores the corresponding formulas for engineering calculations.

[0049] Regular formulas require regular updates as new formulas emerge over time.

[0050] 3) The calculation and comparison module synchronizes the attendance data of the attendance system and the engineering management data of the engineering management system through the data synchronization tool, extracts the attendance data into the accounting module for salary calculation, and extracts the engineering management data into the budget module for engineering calculation.

[0051] The accounting module brings the attendance data into the calculation formulas of various salary operations, and the budget module brings the engineering management data into the calculation formulas of various engineering operations, obtaining different salary calculation results and engineering calculation results, and comparing the different calculation results.

[0052] The synchronization tool is used to synchronize the data of the attendance system, which can avoid manual data entry. The synchronization tool is an existing technology and will not be introduced in detail here.

[0053] 4) The recommendation module is provided with recommendation conditions corresponding to each fund operation, and the calculation formula corresponding to the calculation result that meets the recommendation conditions is recommended.

[0054] Recommended conditions are conditions set according to the needs of the enterprise. For salary calculations, the recommended conditions can be the lowest salary cost or the salary cost within a certain set range; for engineering calculations, the recommended conditions can be the lowest budget or the budget within a certain set range.

[0055] The recommendation conditions can be adjusted according to time factors and operational factors, and the recommendation results will be updated after the recommendation conditions are adjusted. For example, if the company is in difficulty, the recommended condition for salary calculation is set to the lowest salary cost, and if the company's operating conditions are alleviated, the recommended condition for salary calculation is set to moderate salary cost.

[0056] In the above embodiment, the formula definition module is used to build a custom formula in order to meet the work needs of staff at different levels. Even staff with little work experience can complete the input of the formula. As other implementation methods, the formula can also be built in a programming manner, and the present invention does not limit this.

[0057] The following describes in detail the implementation method of the formula recommendation algorithm of the present invention by taking the implementation of the calculation formula recommendation algorithm of the functional position in the salary calculation as an example.

[0058] First, the intelligent formula learning library stores the formula library 001 for functional positions and the formula library 002 for functional positions.

[0059] Functional Position Formula Library 001

[0060] 1. Car fare subsidy: iff(whether to provide car fare subsidy = "1",iff(number of days to be on duty>0,60 / (number of days to be on duty - number of holidays in the month)*total car fare subsidy in the month,0),0);

[0061] 2. Full attendance hours: 8*full attendance days;

[0062] 3. The company bears social insurance: the pension company bears the current month + the medical company bears the current month + the work injury company bears the current month + the childbirth company bears the current month + the unemployment company bears the current month + the housing company bears the current month + the pension company bears the previous month + the housing company bears the previous month;

[0063] 4. Deductions for missed attendance: iff(statutory holidays during the attendance period <> 0 or full attendance days - required attendance days - number of holidays in the month > 0, (basic salary + position salary) / full attendance hours * (full attendance hours - required attendance hours - statutory holidays during the attendance period * 8), 0);

[0064] 5. Overtime pay: basic salary / 21.75 / 8*holiday overtime hours*3;

[0065] 6. Meal allowance: 10*number of overtime meal allowances on rest days + 10*number of meal allowances on working days and night shifts;

[0066] 7. Deductions for leave: iff(duty hours>0, (basic salary + post salary + certificate allowance) / (full attendance hours)*(personal leave hours + sick leave hours*0.5+abortion leave hours*0.5+maternity leave days*8), 0);

[0067] 8. Deductions for absenteeism: iff(number of hours to be attended>0, (basic salary+position salary+certificate allowance) / (number of full attendance hours)*number of absenteeism hours*2,0);

[0068] 9. Pay in lieu of notice: iff(whether to deduct pay in lieu of notice = 1, (basic salary + position salary + meal allowance + transportation allowance) / 30*number of days of notice in lieu of notice, 0);

[0069] 10. Total payable: basic salary + position salary + meal allowance + travel allowance + certificate allowance + overtime pay + overtime allowance + overtime meal allowance + project allowance + process allowance + back pay + other allowances + holiday allowance + workload bonus + performance bonus;

[0070] 11. Total income: total payable - (quality award deductions + full attendance award deductions + lateness deductions + leave deductions + absenteeism deductions + no-attendance deductions + other deductions + supplementary deductions + notice pay);

[0071] 12. Taxable wages: total income - (previous month pension + previous month medical + previous month unemployment + previous month housing + current month pension + current month work injury + current month medical + current month childbirth + current month unemployment + current month housing + current month commercial insurance + next month pension + next month work injury + next month medical + next month childbirth + next month unemployment + next month housing) + cash payment items;

[0072] 13. Wages payable: Total income - (pension for the current month + work injury for the current month + medical treatment for the current month + childbirth for the current month + unemployment for the current month + housing for the current month)

[0073] 14. Total deductions: quality award deductions + full attendance award deductions + lateness deductions + leave deductions + absenteeism deductions + no-attendance deductions + other deductions + back pay + notice pay;

[0074] 15. Total actually paid: Total payable - Total deductions - Total withheld amounts.

[0075] Functional Position Formula Library 002

[0076] 1. Transportation subsidy: iff(whether to send transportation subsidy = "1", 60, 0);

[0077] 2. Full attendance hours: 8*full attendance days;

[0078] 3. Social security borne by the company: pension company for the current month + medical company for the current month + unemployment company for the current month + work-related injury company for the current month + birth company for the current month + housing company for the current month + pension company for the previous month + housing company for the previous month + social security adjustment items + outpatient medical company for the current month + major disease medical company for the previous month + local pension + employment security fund for the disabled + work-related injury company for the previous month + medical company for the previous month + birth company for the previous month + unemployment company for the previous month + supplementary work-related injury insurance + major disease medical company for the previous month; (This formula is applicable to the calculation method of deducting the social security of the current month or deducting the social security of the previous month, choose one of the two);

[0079] 4. Deductions for non-attendance: iff(full attendance days - required attendance days - personal leave days - sick leave days - other leave days - statutory holidays during the attendance period>0,(basic salary + position salary + regional subsidy + housing subsidy + telephone subsidy + car fare subsidy + certificate allowance + process subsidy) / full attendance days*(full attendance days - required attendance days - personal leave days - sick leave days - other leave days - statutory holidays during the attendance period), 0)”;

[0080] 5. Overtime pay: basic salary / 21.75 / 8*3*holiday overtime hours;

[0081] 6. Meal allowance: iff(whether to issue meal allowance = "1", number of days to be worked * 12, 0);

[0082] 7. Deductions for leave: iff(number of personal leave days + number of sick leave days>3, (basic salary + position salary + car fare allowance + regional subsidy + housing allowance + telephone allowance + certificate allowance + process allowance) / number of full attendance days*number of personal leave days+(basic salary + position salary + regional subsidy + housing allowance + telephone allowance + car fare allowance + certificate allowance + process allowance) / number of full attendance days*number of sick leave days*0.5,0);

[0083] 8. Pay in lieu of notice: iff(whether to deduct pay in lieu of notice = 1, (basic salary + position salary + regional subsidy + meal allowance + transportation allowance + housing allowance + telephone allowance + certificate allowance) / 30*number of days of notice in lieu of notice, 0);

[0084] 9. Total payable: basic salary + position salary + regional subsidy + meal allowance + transportation allowance + certificate allowance + housing allowance + process allowance + telephone allowance + overtime pay + project allowance + holiday allowance + back pay + other allowances + performance bonus + workload bonus;

[0085] 10. Total income: total payable - (full attendance bonus deduction + quality bonus deduction + leave deduction + absence deduction + other deductions + notice pay + supplementary deductions);

[0086] 11. Taxable wages: total income - (pension in the previous month + medical treatment in the previous month + major illness medical treatment in the previous month + unemployment in the previous month + housing in the previous month);

[0087] 12. Total deductions: full attendance award deductions + quality award deductions + leave deductions + absence deductions + other deductions + notice pay + supplementary deductions for pension + medical care + unemployment + housing + outpatient medical care withholding tax + work-related injury + childbirth;

[0088] 13. Wages payable: total income - (pension for the current month + medical insurance for the current month + unemployment for the current month + housing for the current month + work-related injury for the current month + commercial insurance for the current month + childbirth for the current month);

[0089] 14. Total actually paid: Total payable - Total deductions - Total withheld amounts.

[0090] Secondly, the calculation and comparison module performs data synchronization or import, calculation of the formula library and compares the calculation results.

[0091] Used for attendance personnel information, salary information, and attendance data. The above data sources will be extracted from the business system through the data synchronization service, such as: attendance records, leave records, overtime records, absenteeism records, holiday records, etc. from the attendance system.

[0092] The basic salary in the data source warehouse is: 4,000 yuan; the position salary is: 6,000 yuan; the number of full attendance days is: 22 days; the actual number of attendance days is: 21 days; the overtime time is 15 hours; the regional subsidy is: 100 yuan; the housing subsidy is: 1,000 yuan; the telephone subsidy is 100 yuan; the certificate allowance is 100 yuan; the process subsidy is 100 yuan; the number of overtime meal allowances on rest days: 10 times, the number of meal allowances for working days and night shifts: 10 times; the number of days to be notified is: 0 days; the number of statutory holidays is: 0 days; the monthly pension is: 360 yuan; the monthly medical insurance is: 90 yuan; the monthly unemployment insurance is: 90 yuan; the monthly housing is: 450 yuan;

[0093] Bring the above data into the functional position formula library 001 and the functional position formula library 002, and the calculation results of the functional position 001 and 002 formula libraries are as follows:

[0094] Calculation results of functional position formula library 001

[0095] 1. Transportation subsidy: 40 yuan;

[0096] 2. Full attendance hours: 8*22=176 hours;

[0097] 3. The company bears social security: 360+90+13.5+450=913.5 yuan;

[0098] 4. Deduction for absence from work: (4000+6000) / 22*1=454.55 yuan;

[0099] 5. Overtime pay: 4000 / 21.75 / 8*15*3=1034.48 yuan;

[0100] 6. Meal allowance: 10*10+10*10=200 yuan;

[0101] 7. Leave deduction: (4000+6000+100) / (176)*8=459 yuan;

[0102] 8. Deduction for absenteeism: 0 yuan;

[0103] 9. Notice fee: (4000+40+200) / 30*0=0 yuan;

[0104] 10.Total payable:

[0105] 4000+6000+100+200+40+100+1034.48+0+0+0+100+0+0+0+0+0+0=11474.48 yuan;

[0106] 11. Total income: 11474.48-(0+0+0+0+0+454.55+0+0+0)=11019.93 yuan;

[0107] 12. Taxable wages: 11019.93-(360+90+13.5+450)=10106.43 yuan;

[0108] 13. Salary payable: 11019.93-(360+90+13.5+450)=10106.43 yuan;

[0109] 14. Total deductions: 0+0+0+0+0+454.55+0+0+0=454.55 yuan;

[0110] 15. Actual total amount paid: 11474.48-454.55-0=11019.93 yuan.

[0111] Calculation results of functional position formula library 002

[0112] 1. Transportation subsidy: 60 yuan;

[0113] 2. Full attendance hours: 8*22=176 hours;

[0114] 3. The company bears social insurance:

[0115] 360+90+13.5+450+100+1000+0+0+0+0+0+0+0+0+0=2013.5 yuan;

[0116] 4. Deductions for missed attendance: (4000+6000+100+1000+100+100+100) / 22*1=518.18 yuan;

[0117] 5. Overtime pay: 4000 / 21.75 / 8*15*3=1034.48 yuan;

[0118] 6. Meal allowance: 22*12=264 yuan;

[0119] 7. Deductions for leave:

[0120] (4000+6000+60+100+1000+100+100+100) / 22*1+(4000+6000+100+1000+100+60+100+100) / 22*0*0.5,0)=520 yuan;

[0121] 8. Notice fee: (4000+100+264+60+100+100+100) / 30*0=0 yuan;

[0122] 9.Total payable:

[0123] 4000+6000+100+264+60+100+1000+100+100+1034.48+0+0+0+0+0+0=12758.48 yuan;

[0124] 10. Total income: 12758.48-(0+0+0+518.18+0+0+0)=12240.3 yuan;

[0125] 11. Taxable wages: 12240.3-(360+90+13.5+450)=11326.8 yuan;

[0126] 12. Total deductions: 0+0+0+518.18+0+0+0=518.18 yuan;

[0127] 13. Salary payable: 12240.3-(360+90+13.5+450)=11326.8 yuan;

[0128] 14. Actual total amount paid: 12758.48-518.18-0=12240.3 yuan.

[0129] The calculation comparison module automatically compares the actual salary calculation result of functional position formula library 001, which is 11,019.93 yuan, and the actual salary calculation result of functional position formula library 002 is 12,240.3 yuan, and so on. The system will automatically calculate the salary calculation results of all formula libraries in the [Intelligent Formula Learning Library] (including functional position formula library 003, functional position formula library 004...), and sort the calculation results from large to small after automatic comparison.

[0130] Finally, the recommendation module customizes the input conditions (i.e., recommendation conditions) to obtain intelligent formula recommendation results.

[0131] Customize input conditions, such as recommending the formula with the lowest salary cost.

[0132] The system automatically compares all calculation results and intelligently recommends the following result: Functional Position Formula Library 001 is the formula with the lowest salary cost for functional positions.

[0133] It can be seen from the above specific embodiments that the salary calculation process is very complicated and the efficiency of manual learning is very low. The present invention obtains the optimal recommended formula by automatically comparing different calculation formulas, thereby greatly improving the efficiency of selecting the optimal formula.

[0134] The main application scenarios of the formula recommendation of the present invention are: salary settlement, project budget, etc., and are not limited to these two types of financial operations. The present invention provides the best and most suitable algorithm, provides the best solution for enterprises, and reduces the cost increase caused by various trial and error costs. Take the customer we provide this product to as an example: if a manufacturing factory with 3,000 people, with an average of 25 yuan per working hour, using the attendance recommendation algorithm saves nearly 75,000 yuan per month compared to before using the algorithm.

[0135] Implementation method of formula recommendation algorithm:

[0136] The specific implementation process and effects of the implementation method of the formula recommendation algorithm have been introduced in the above-mentioned implementation system embodiment of the formula recommendation algorithm, and will not be repeated here.

Claims

1. A method for implementing a formula recommendation algorithm, characterized in that: The following steps are involved: 1) Pre-store several calculation formulas for fund operations, including conventional formulas and / or custom formulas; conventional formulas are unified calculation formulas developed based on expert experience; custom formulas are calculation formulas built based on demand; fund operations include salary operations and / or engineering operations; 2) Obtain the data source corresponding to each fund operation and determine the corresponding recommendation conditions; the data source corresponding to the salary operation is attendance data, and the data source corresponding to the engineering operation is engineering management data; Recommended conditions are conditions set according to requirements; 3) Bring the data source into the corresponding different calculation formulas for fund operation and compare the calculation results of different calculation formulas; 4) Recommend the calculation formula corresponding to the calculation results that meet the recommended conditions; The recommended conditions are set according to the needs of the enterprise. For salary calculation, the recommended condition is the lowest salary cost; for engineering calculation, the recommended condition is the lowest budget; Among them, the recommendation conditions are adjusted according to time factors and operational factors, and the recommendation results are updated after the recommendation conditions are adjusted.

2. The method for implementing the formula recommendation algorithm according to claim 1, characterized in that: Custom formulas are built through the formula definition module, which includes operators and functions; operators include arithmetic operators, comparison operators, and logical operators; functions include sine function, cosine function, tangent function, absolute value function, and judgment function.

3. The method for implementing the formula recommendation algorithm according to claim 1, characterized in that: The general formula is updated regularly.

4. A system for implementing a formula recommendation algorithm, characterized in that: It includes an intelligent formula learning library, a calculation comparison module and a recommendation module. The intelligent formula learning library is used to store several calculation formulas for fund operations; the several calculation formulas include conventional formulas and / or custom formulas; conventional formulas are unified calculation formulas formulated based on expert experience; custom formulas are calculation formulas built by oneself according to needs; fund operations include salary operations and / or engineering operations; the calculation comparison module is used to obtain the data source corresponding to each fund operation; Bring the data source into the different calculation formulas of the corresponding capital operation and compare the calculation results of different calculation formulas; the data source corresponding to the salary operation is the attendance data, and the data source corresponding to the engineering operation is the engineering management data; The recommendation module is used to determine the recommendation conditions corresponding to each fund operation, and recommend the calculation formula corresponding to the calculation result that meets the recommendation conditions; Recommended conditions are conditions set according to requirements; The recommended conditions are set according to the needs of the enterprise. For salary calculation, the recommended condition is the lowest salary cost; for engineering calculation, the recommended condition is the lowest budget; The recommendation conditions are adjusted based on time factors and operational factors, and the recommendation results are updated after the recommendation conditions are adjusted.

5. The implementation system of the formula recommendation algorithm according to claim 4, characterized in that: It also includes a formula definition module. Custom formulas are built through the formula definition module. The formula definition module includes operators and functions; operators include arithmetic operators, comparison operators and logical operators; functions include sine function, cosine function, tangent function, absolute value function and judgment function.

6. The implementation system of the formula recommendation algorithm according to claim 4, characterized in that: The general formula is updated regularly.

Images