Car insurance customer loss prediction method based on big data

Abstract

The invention relates to the technical field of car insurance, and discloses a big data-based car insurance customer loss prediction method, which comprises the steps of establishing a user driving behavior set, calculating a user driving behavior evaluation index according to the user driving behavior set, and establishing a user insurance information set, calculating a user insurance risk index according to the user insurance information set, establishing a user information interaction set, calculating a user interaction satisfaction degree according to the user information interaction set, and calculating a user churn degree in combination with user driving behavior evaluation, the user insurance risk index and the user interaction satisfaction degree. Vehicle insurance customer loss prediction is carried out according to the user loss degree, feedback is carried out for lost customers, and the actual conditions and demands of the customers can be more comprehensively known by integrating multi-dimensional data such as user driving behaviors, insurance risk indexes and interaction satisfaction, so that the customer loss prediction accuracy is improved, and the user experience is improved. And the enterprise can quickly identify the user with relatively high loss risk and perform intervention.

Description

Technical Field

[0001] The present invention relates to the technical field of auto insurance, and in particular to an auto insurance customer churn prediction method based on big data. Background Art

[0002] Auto insurance, usually refers to motor vehicle insurance, is a commercial insurance that is responsible for compensating for personal injury or property damage caused by natural disasters or accidents involving motor vehicles. Auto insurance is a type of transportation insurance that provides protection against motor vehicles and their third-party liabilities. It mainly includes corporate entities and individuals who own various motor vehicles, including various types of cars and special vehicles such as trams, electric vehicles, and motorcycles.

[0003] Auto insurance business is one of the important sources of income for insurance companies. With the increase in the number of cars, the auto insurance market continues to expand, providing insurance companies with stable premium income. If the loss of auto insurance customers is serious, it will directly lead to a reduction in the insurance company's premium income, which will put pressure on the company's financial situation. In a highly competitive market, customer loss may cause the insurance company's market share to gradually decline. As more and more customers turn to other competitors, the company's position in the market may be shaken. Existing insurance companies are unable to accurately predict the loss of auto insurance customers, resulting in serious customer loss. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In view of the shortcomings of the prior art, the present invention provides a method for predicting auto insurance customer churn based on big data. By integrating multi-dimensional data such as user driving behavior, insurance risk index and interaction satisfaction, the method can more comprehensively understand the actual situation and needs of customers, thereby improving the accuracy of customer churn prediction and reducing business losses caused by customer churn.

[0006] (II) Technical solution

[0007] To achieve the above object, the present invention provides the following technical solution: a method for predicting auto insurance customer churn based on big data, comprising the following steps:

[0008] Step 1: Establish a user driving behavior set;

[0009] Step 2: Calculate the user driving behavior evaluation index based on the user driving behavior set;

[0010] Step 3: Establish user insurance information collection;

[0011] Step 4: Calculate the user insurance risk index based on the user insurance information set;

[0012] Step 5: Establish a user information interaction set;

[0013] Step 6: Calculate user interaction satisfaction based on the user information interaction set;

[0014] Step 7: Calculate user churn by combining user driving behavior evaluation, user insurance risk index, and user interaction satisfaction;

[0015] Step 8. Predict auto insurance customer churn based on user churn rate and provide feedback to churned customers.

[0016] Preferably, in step one, the user driving behavior set is obtained through Internet of Vehicles equipment, GPS and other technologies, and the user driving behavior set includes the average vehicle speed data, acceleration frequency, total driving time, lane change frequency, overtaking frequency, braking frequency, and average driving time of the auto insurance user.

[0017] Preferably, in step 2, the calculation formula of the user driving behavior evaluation index is:

[0018]

[0019] In the above formula, JSpg is the user driving behavior evaluation index, Pc i is the user’s average vehicle speed data, Pc o is the ideal average vehicle speed data, Jp is the acceleration frequency, ZJ is the total driving time, Bp is the lane change frequency, Cp is the overtaking frequency, Sp is the braking frequency, and Jp is the average driving time;

[0020] ω1, ω2, ω3, ω4, ω5, and ω6 are corresponding weight factors, which are used to adjust the contribution of each factor to the user's driving behavior evaluation index, and ω1+ω2+ω3+ω4+ω5+ω6=1;

[0021] Indicates the percentage difference between the average vehicle speed data and the ideal average vehicle speed data;

[0022] It represents the logarithm of the ratio of acceleration frequency to total driving time;

[0023] Represents the logarithm of the ratio of lane change frequency to total driving time;

[0024] represents the logarithm of the ratio of overtaking frequency to total driving time;

[0025] Represents the logarithm of the ratio of braking frequency to total driving time.

[0026] Preferably, in step three, the user insurance information set includes the user's historical number of accidents, payment delay rate, accident loss amount, accident claim amount, and claim processing efficiency.

[0027] Preferably, in step 4, the calculation formula of the user insurance risk index is:

[0028]

[0029] In the above formula, BFpg is the user's insurance risk index, Cc is the number of historical accidents of the user, Jy is the payment delay rate, Sj is the i is the amount of loss caused by the accident, Sj max is the maximum amount of loss caused by the accident, Lj i is the amount of compensation for the accident, Lj max is the maximum amount of claims, and Cx is the efficiency of claims processing;

[0030] Represents the ratio of the insured loss amount to the maximum loss amount;

[0031] Represents the ratio of the claim amount to the maximum claim amount;

[0032] 1-Cx represents the value opposite to the claims processing efficiency;

[0033] ρ1, ρ2, ρ3, ρ4, and ρ5 are weight factors of the corresponding factors respectively, and ρ1+ρ2+ρ3+ρ4+ρ5=1.

[0034] Preferably, in step five, the user information interaction set includes the frequency of users contacting customer service, the user's historical ratings, the frequency of users' intention to change their insurance policies, and the degree of participation in value-added activities.

[0035] Preferably, in step 6, the calculation formula for user interaction satisfaction is:

[0036]

[0037] In the above formula, JHmy is the user interaction satisfaction, Kp is the frequency of users contacting customer service, Lp is the user's historical score, Gp is ​​the frequency of users' intention to change the insurance policy, Zd is the participation in value-added activities, θ1, θ2, θ3, θ4 are the corresponding weight factors, and θ1+θ2+θ3+θ4=1.

[0038] Preferably, in step seven, the user churn rate is calculated as follows:

[0039] YHls=JSpg*BFpg*JHmy

[0040] In the above formula, YHls represents the user churn rate.

[0041] Preferably, when the user churn degree YHls is less than the user churn standard value, it means that the churn degree of this user is relatively high.

[0042] Preferably, a churn customer attention report is generated for the customers with a higher churn degree, and focused feedback and attention are given to the customers with a higher churn degree.

[0043] Compared with the prior art, the present invention provides a method for predicting auto insurance customer churn based on big data, which has the following beneficial effects:

[0044] 1. The present invention uses the user's driving behavior data to enable insurance companies to more accurately assess risks and thus formulate personalized premiums. Insurance companies can identify high-risk drivers and take preventive measures to reduce accident rates and claims costs. Through the user's insurance risk index, the auto insurance customer churn prediction can be used to more accurately identify high-risk customers and take more stringent risk management measures for these high-risk customers. Through user interaction satisfaction, customers who are dissatisfied with the service can be more accurately identified, and insurance companies can take timely measures, such as improving services or providing compensation, to prevent the loss of these customers. Insurance companies can discover weak links in services and make targeted improvements to improve overall service quality, enhance customer experience, and reduce customer churn due to service dissatisfaction.

[0045] 2. By integrating multi-dimensional data such as user driving behavior, insurance risk index and interaction satisfaction, the present invention can more comprehensively understand the actual situation and needs of customers, thereby improving the accuracy of customer churn prediction, and can issue early warnings before customer churn, so that companies have enough time to take intervention measures to reduce customer churn rates. Effective customer churn predictions and intervention measures can significantly improve customer retention rates and reduce business losses caused by customer churn. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a step diagram of the method of the present invention; DETAILED DESCRIPTION

[0047] 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.

[0048] See also Figure 1 , a method for predicting auto insurance customer churn based on big data, comprising the following steps:

[0049] Step 1: Establish a user driving behavior set;

[0050] The user driving behavior set is obtained through Internet of Vehicles devices, GPS and other technologies. The user driving behavior set includes the average speed data, acceleration frequency, total driving time, lane change frequency, overtaking frequency, braking frequency, and average driving time of the auto insurance user;

[0051] Step 2: Calculate the user driving behavior evaluation index based on the user driving behavior set;

[0052] The calculation formula of the user driving behavior evaluation index is:

[0053]

[0054] In the above formula, JSpg is the user driving behavior evaluation index, Pc i is the user’s average vehicle speed data, Pc o is the ideal average vehicle speed data, Jp is the acceleration frequency, which indicates the frequency of acceleration of the user during driving, ZJ is the total driving time, which indicates the total driving time of the user, Bp is the lane change frequency, which indicates the number or frequency of lane changes of the user during driving, Cp is the overtaking frequency, which indicates the number or frequency of overtaking of the user during driving, Sp is the braking frequency, which indicates the number or frequency of braking of the user during driving, and Jp is the average driving time;

[0055] ω1, ω2, ω3, ω4, ω5, and ω6 are corresponding weight factors, which are used to adjust the contribution of each factor to the user's driving behavior evaluation index, and ω1+ω2+ω3+ω4+ω5+ω6=1;

[0056] Indicates the percentage difference between the average speed data and the ideal average speed data, reflecting whether the user's speed is close to the ideal speed;

[0057] It represents the logarithm of the ratio of acceleration frequency to total driving time, which measures the frequency of acceleration during driving.

[0058] It represents the logarithm of the ratio of lane change frequency to total driving time, measuring the frequency of lane changes during driving.

[0059] It represents the logarithm of the ratio of overtaking frequency to total driving time, measuring the frequency of overtaking by the user during driving;

[0060] It represents the logarithm of the ratio of braking frequency to total driving time, which measures the frequency of braking during driving.

[0061] Based on the user's driving behavior data, insurance companies can more accurately assess risks and formulate personalized premiums. Insurance companies can identify high-risk drivers and take preventive measures to reduce accident rates and claims costs, while improving customer safety and satisfaction.

[0062] Step 3: Establish user insurance information collection;

[0063] The user insurance information set includes the number of historical accidents, payment delay rate, accident loss amount, accident claim amount, and claim processing efficiency;

[0064] Step 4: Calculate the user insurance risk index based on the user insurance information set;

[0065] The calculation formula of user insurance risk index is:

[0066]

[0067] In the above formula, BFpg is the user's insurance risk index, Cc is the number of historical accidents of the user, which is an indicator used to measure the frequency of accidents that occurred to the user in the past. A higher number of accidents may mean a higher future risk, Jy is the payment delay rate, which indicates whether the user pays the premium on time. Delayed payment may indicate problems with the user's financial stability, thereby increasing risk, and Sj i is the amount of loss caused by the accident, Sj max is the maximum amount of loss caused by the accident, Lj i is the amount of compensation for the accident, Lj max is the maximum amount of claims, and Cx is the efficiency of claims processing;

[0068] Represents the ratio of the insured loss amount to the maximum loss amount, which is used to standardize the insured loss amount so as to compare it with the insured losses of other users;

[0069] Represents the ratio of the amount of claims paid to the maximum claim amount, and is used to standardize claim amounts to assess the relative size of claims;

[0070] 1-Cx represents the opposite value of claims processing efficiency, which is used to indicate the inefficiency of claims processing efficiency. If it is close to 1, it means that the claims processing efficiency is high and the risk is low. If it is close to 0, it means that the claims processing efficiency is low and the risk is high.

[0071] ρ1, ρ2, ρ3, ρ4, and ρ5 are weight factors of the corresponding factors, and ρ1+ρ2+ρ3+ρ4+ρ5=1;

[0072] Through the user insurance risk index, the prediction of auto insurance customer churn can more accurately identify high-risk customers and take more stringent risk management measures for these high-risk customers. Insurance companies can identify customers who may face higher risks in advance and take preventive measures to enhance customer experience, improve customer loyalty, and reduce customer churn.

[0073] Step 5: Establish a user information interaction set;

[0074] The user information interaction set includes the frequency of users contacting customer service, the user's historical rating, the frequency of users' intention to change their insurance policies, and the degree of participation in value-added activities;

[0075] Step 6: Calculate user interaction satisfaction based on the user information interaction set;

[0076] The calculation formula for user interaction satisfaction is:

[0077]

[0078] In the above formula, JHmy is the user interaction satisfaction, Kp is the frequency of users contacting customer service, which usually reflects the user's dependence on and satisfaction with the service. Frequent contact may mean that the user has encountered a problem or needs additional help, which may be due to complex service processes, inconvenient product use, or users have questions about certain terms. Lp is the user's historical score, which directly reflects the user's satisfaction with the insurance company's past services. Gp is ​​the frequency of users' intention to change their policies, which can reveal the user's adaptability and satisfaction with the current insurance product. Zd is the participation in value-added activities, which reflects the user's interest in and willingness to participate in additional services or activities provided by the insurance company. θ1, θ2, θ3, θ4 are corresponding weight factors, and θ1+θ2+θ3+θ4=1;

[0079] Customers who are dissatisfied with the service can be more accurately identified through user interaction satisfaction. Insurance companies can take timely measures, such as improving services or providing compensation, to prevent these customers from churn. Insurance companies can find weak links in services and make targeted improvements to improve overall service quality, enhance customer experience, and reduce customer churn due to service dissatisfaction.

[0080] Step 7: Calculate user churn by combining user driving behavior evaluation, user insurance risk index, and user interaction satisfaction;

[0081] The calculation method of user churn is:

[0082] YHls=JSpg*BFpg*JHmy

[0083] In the above formula, YHls represents the user churn degree;

[0084] Step 8: Predict auto insurance customer churn based on user churn rate, and provide feedback to churned customers;

[0085] When the user churn degree YHls is less than the user churn standard value, it means that the churn degree of this user is high;

[0086] Generate a churn customer attention report for customers with a high churn rate, and provide key feedback and attention to customers with a high churn rate;

[0087] By integrating multi-dimensional data such as user driving behavior, insurance risk index and interaction satisfaction, we can have a more comprehensive understanding of the actual situation and needs of customers, thereby improving the accuracy of customer churn predictions. We can issue early warnings before customer churn occurs, giving companies enough time to take intervention measures to reduce customer churn rates. Effective customer churn predictions and intervention measures can significantly improve customer retention rates and reduce business losses caused by customer churn.

[0088] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for predicting auto insurance customer churn based on big data, characterized in that: The following steps are involved: Step 1: Establish a user driving behavior set; Step 2: Calculate the user driving behavior evaluation index based on the user driving behavior set; Step 3: Establish user insurance information collection; Step 4: Calculate the user insurance risk index based on the user insurance information set; Step 5: Establish a user information interaction set; Step 6: Calculate user interaction satisfaction based on the user information interaction set; Step 7: Calculate user churn by combining user driving behavior evaluation, user insurance risk index, and user interaction satisfaction; Step 8. Predict auto insurance customer churn based on user churn rate and provide feedback to churned customers.

2. The method for predicting auto insurance customer churn based on big data according to claim 1, characterized in that: In step one, the user driving behavior set is obtained through Internet of Vehicles devices, GPS and other technologies. The user driving behavior set includes the average vehicle speed data, acceleration frequency, total driving time, lane change frequency, overtaking frequency, braking frequency, and average driving time of the auto insurance user.

3. The method for predicting auto insurance customer churn based on big data according to claim 2, characterized in that: In step 2, the calculation formula of the user driving behavior evaluation index is: In the above formula, JSpg is the user driving behavior evaluation index, Pc i is the user’s average vehicle speed data, Pc o is the ideal average vehicle speed data, Jp is the acceleration frequency, ZJ is the total driving time, Bp is the lane change frequency, Cp is the overtaking frequency, Sp is the braking frequency, and Jp is the average driving time; ω1, ω2, ω3, ω4, ω5, and ω6 are corresponding weight factors, which are used to adjust the contribution of each factor to the user's driving behavior evaluation index, and ω1+ω2+ω3+ω4+ω5+ω6=1; Indicates the percentage difference between the average vehicle speed data and the ideal average vehicle speed data; It represents the logarithm of the ratio of acceleration frequency to total driving time; Represents the logarithm of the ratio of lane change frequency to total driving time; represents the logarithm of the ratio of overtaking frequency to total driving time; Represents the logarithm of the ratio of braking frequency to total driving time.

4. The method for predicting auto insurance customer churn based on big data according to claim 1, characterized in that: In the step three, the user insurance information set includes the user's historical number of accidents, payment delay rate, accident loss amount, accident claim amount, and claim processing efficiency.

5. The method for predicting auto insurance customer churn based on big data according to claim 4, characterized in that: In step 4, the calculation formula of the user insurance risk index is: In the above formula, BFpg is the user's insurance risk index, Cc is the number of historical accidents of the user, Jy is the payment delay rate, Sj is the i is the amount of loss caused by the accident, Sj max is the maximum amount of loss caused by the accident, Lj i is the amount of compensation for the accident, Lj max is the maximum amount of claims, and Cx is the efficiency of claims processing; Represents the ratio of the insured loss amount to the maximum loss amount; Represents the ratio of the claim amount to the maximum claim amount; 1-Cx represents the value opposite to the claims processing efficiency; ρ1, ρ2, ρ3, ρ4, and ρ5 are weight factors of the corresponding factors respectively, and ρ1+ρ2+ρ3+ρ4+ρ5=1.

6. The method for predicting auto insurance customer churn based on big data according to claim 1, characterized in that: In step five, the user information interaction set includes the frequency of users contacting customer service, the user's historical ratings, the frequency of users' intention to change their insurance policies, and the degree of participation in value-added activities.

7. The method for predicting auto insurance customer churn based on big data according to claim 6, characterized in that: In step 6, the calculation formula of user interaction satisfaction is: In the above formula, JHmy is the user interaction satisfaction, Kp is the frequency of users contacting customer service, Lp is the user's historical score, Gp is ​​the frequency of users' intention to change their insurance policies, and Zd is the participation in value-added activities. is the corresponding weight factor, and 8. The method for predicting auto insurance customer churn based on big data according to claim 7, characterized in that: In step 7, the user churn rate is calculated as follows: YHls=JSpg*BFpg*JHmy In the above formula, YHls represents the user churn rate.

9. The method for predicting auto insurance customer churn based on big data according to claim 8, characterized in that: When the user churn degree YHls is less than the user churn standard value, it means that the churn degree of this user is high.

10. The method for predicting auto insurance customer churn based on big data according to claim 9, characterized in that: A churned customer attention report is generated for the customers with a high churn degree, and key feedback and attention are given to the customers with a high churn degree.

Images