A method for improving vehicle rescue success rate based on big data

Abstract

The application discloses a method for improving vehicle rescue success rate based on big data, which comprises the following steps: obtaining all basic tags as enumeration factors through a road rescue big data system tag library; classifying the enumeration factors into basic factors, special factors and adjustment factors through a COX regression model, and respectively assigning values to the basic factors, the special factors and the adjustment factors; a user fills in user information and user rescue information; according to the user information and the user rescue information, the most suitable rescue order receiving personnel is matched through a call matching algorithm, and rescue demand information is accurately pushed; after receiving the rescue demand information, the rescue order receiving personnel confirms whether to accept the order by himself / herself and goes to the accident point for rescue. Through the call matching algorithm, all users who are willing to volunteer for rescue in the surrounding area can be matched, the order receiving range of vehicle rescue demand is expanded, the most suitable vehicle rescue personnel is matched, and the success rate of vehicle rescue is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vehicle rescue, in particular to a method for improving the success rate of vehicle rescue based on big data. BACKGROUND

[0002] Road rescue refers to automobile road emergency rescue, which provides on-site repair services for the owner of the faulty vehicle. It also refers to road rescue in traffic accidents, including first aid for the injured and road dredging. With the continuous development of the domestic automobile market, large and medium-sized automobile maintenance and service companies have launched their own automobile rescue services. If the owner encounters a breakdown, such as sudden failure to start, failure to start after the engine stalls, no oil, no electricity, or even a flat tire on the way, they can contact the 4S store or the nearest automobile rescue service company and seek help from professional automobile rescue technicians. They will ask and determine the approximate problem of the car and then go to the scene to provide assistance. General automobile rescue services include: oil delivery, charging, tire replacement, on-site troubleshooting, fault towing, and on-site rescue guidance.

[0003] The service objects of automobile rescue (vehicle rescue) include: novice car owners, who may not be familiar with cars and are likely to be confused when the car suddenly fails to move. Road rescue services are very important at this time. Friends who love self-driving tours, who are afraid of unexpected breakdowns on the way, and who do not rely on villages or stores. Calling a tow truck service directly is very expensive. All cars that are not within the warranty period, 4S stores no longer have the obligation to provide free rescue services for cars that are not within the warranty period. Cars that do not provide rescue services by the car brand, some low-end cars do not provide any rescue services, and these car owners really need a driving guarantee.

[0004] Currently, in the field of vehicle rescue, when a user raises a vehicle rescue demand, only insurance companies or vehicle brand merchants are supported to perform rescue operations by professional teams, which has many problems: such as long road rescue order time, low user satisfaction rate, and unclear road rescue order information. The original road rescue system has no precise matching algorithm, which causes difficulty in road rescue, increases the cost of road rescue, and increases the difficulty and time of road rescue. SUMMARY

[0005] The application aims to provide a method for improving vehicle rescue success rate based on big data, so as to solve the technical problems of narrow road rescue business scope, limitation to insurance and brand companies, long user road rescue order receiving time and low user satisfaction rate.

[0006] The application aims to achieve the above technical problems by the following technical scheme: a method for improving vehicle rescue success rate based on big data, comprising the following steps:

[0007] Factor selection: all basic tags are obtained from the road rescue big data system tag library as enumeration factors;

[0008] Factor classification: the enumeration factors are classified by a COX regression model into basic factors, special factors and adjustment factors, and the basic factors, special factors and adjustment factors are respectively assigned values;

[0009] Initiating rescue demand: a user fills in user information and user rescue information;

[0010] Matching rescue order receiving personnel: the most suitable rescue order receiving personnel are matched by a calling matching algorithm according to the user information and user rescue information, and rescue demand information is accurately pushed;

[0011] Confirming order receiving: after receiving the rescue demand information, the rescue order receiving personnel confirm whether to receive the order and go to the accident point for rescue.

[0012] Further, the formula of the COX regression model is:

[0013] h(t, X) = h0(t)exp(β1X1+β2X2+…+β m X m )

[0014] In the formula, h(t, X) is a hazard rate function, h0(t) is a baseline hazard rate of h(t, X) when the X vector is 0, β1, β2, …, β m are partial regression coefficients of the independent variable X.

[0015] Further, the basic factors include accident type, vehicle attribute, accident level, vehicle brand, vehicle configuration, vehicle type, fuel model, surrounding temperature, weather, season and road condition information; the special factors include distance factor, driver population inclination industry, driver population inclination gender and driving behavior habit; and the adjustment factor includes rescue order receiving personnel praise degree tag.

[0016] Further, the accident types include anchor throwing, no electricity, emergency braking, tire, cosmetic damage, oil tank gauge display, abnormal sound and central control operation; the accident levels include major, large, medium and small; the vehicle brands include high-end cars, medium-end cars and low-end cars; the vehicle configurations include high-end cars, medium-end cars and low-end cars; the vehicle types include fuel cars and electric cars; the fuel types include no fuel, 98# fuel, 95# fuel, 92# fuel and diesel; the weather includes extreme climate, ordinary rain and snow and no special; the seasons include spring, summer, autumn and winter; and the road conditions include mountain, city, suburb, highway or expressway, country road, track and tunnel.

[0017] Further, the formula of the call matching algorithm is:

[0018] SF = log∑[∑a1*h1+∑a2h2+…+∑a n h n ]*b1*k1*b2*k2*…*b m *k m +Y

[0019] In the formula, SF is the matching score of the rescue order taker, a1, a2, …, a n is the basic factor, h1, h2, …, h n is the matching weight of the corresponding basic factor, b1, b2, …, b m is the special factor, k1, k2, …, k m is the matching weight of the corresponding special factor, and Y is the adjustment factor.

[0020] Further, the calculation formula of the basic factor matching weight h n and the special factor matching weight k m is:

[0021] Matching weight = number of orders matched by the rescue order taker corresponding to the factor / total number of orders of the rescue order taker.

[0022] Further, the top N most suitable rescue order takers are matched and locked by the call matching algorithm, and the rescue demand information is accurately distributed to the top N most suitable rescue order takers.

[0023] Further, the order confirmation includes the following sub-steps:

[0024] If someone accepts the order within M minutes, the rescue order matching is successful, and if no one accepts the order, the rescue order takers are diffused according to the calculation results of the call matching algorithm;

[0025] The rescue order takers who receive the rescue demand information determine the rescue order takers who successfully accept the order according to the order acceptance speed;

[0026] If the order in the breach of contract state appears, according to real-time information, the most suitable rescue order matching personnel is urgently matched, and the user is compensated.

[0027] Further, the rescue order matching personnel includes official rescue personnel and ordinary users.

[0028] The beneficial effects of the present application are that: the present application can match all the users (including official rescue personnel and ordinary users) who can volunteer to rescue in the surrounding area through the call matching algorithm, expand the order matching range of vehicle rescue demand, match the most suitable vehicle rescue personnel, improve the success rate of vehicle rescue, and thus improve the satisfaction of vehicle after-sales users; the present application expands the original business of merchants-users into the business of (merchants+users)-users, changes the overall logic of the original business, and also increases the social attribute. BRIEF DESCRIPTION OF DRAWINGS

[0029] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description only show some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from the structures shown in the drawings without creative labor.

[0030] Fig. 1 The flowchart of the present application;

[0031] Fig. 2 The flowchart of the rescue order matching personnel confirming the order;

[0032] Fig. 3 The flowchart of the present application; DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solutions and advantages of the embodiments of the present application more clear, the following will combine the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are some of the embodiments of the present application, not all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations.

[0034] It should be noted that: similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings.

[0035] The following will combine the drawings to make a detailed description of some embodiments of the present application. In the case of no conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0036] Embodiment 1

[0037] Referring to Figs. 1 to 3 A method for improving the success rate of vehicle rescue based on big data, comprising the following steps:

[0038] Factor selection: all basic tags are obtained from the road rescue big data system tag library as enumeration factors;

[0039] Factor classification: enumeration factors are classified by COX regression model, classified as basic factors, special factors and adjustment factors, and respectively assigned to basic factors, special factors and adjustment factors;

[0040] Initiating rescue demand: the user fills in the user information and user rescue information;

[0041] Matching rescue order personnel: according to the user information and user rescue information, the most suitable rescue order personnel is matched through the call matching algorithm, and the rescue demand information is accurately pushed;

[0042] Confirm order: after receiving the rescue demand information, the rescue order personnel confirms whether to accept the order and goes to the accident point for rescue.

[0043] Further, the road rescue big data system tag library is an existing tag library, and at least 100 tags are extracted from the tag library as enumeration factors.

[0044] In this embodiment, the formula of the COX regression model is:

[0045] h(t, X) = h0(t)exp(β1X1+β2X2+…+β m X m ) (1)

[0046] In the formula, h(t, X) is the hazard rate function, h0(t) is the baseline hazard rate of h(t, X) when the X vector is 0, β1, β2, …, β m are the partial regression coefficients of the independent variable X.

[0047] Further, COX regression model, also known as "proportional hazards model" (Cox model for short), is a semi-parametric regression model proposed by British statistician D. R. Cox (1972). The model takes survival outcome and survival time as dependent variable, can analyze the impact of many factors on survival period, can analyze data with censored survival time, and does not require estimation of the survival distribution type of data. Due to the above excellent properties, since the model was introduced, it has been widely used in medical follow-up studies, and is the most widely used multi-factor analysis method in survival analysis. Since COX regression model does not make any assumption on h0(t), COX regression model has great flexibility in handling problems; on the other hand, in many cases, we only need to estimate the parameter β m (such as factor analysis), even in unknown cases, the parameter β m can still be estimated. That is, Cox regression model contains h0(t), so it is not a complete parametric model, but the parameter β m can still be estimated according to formula (1), so COX regression model belongs to semi-parametric model.

[0048] Further, formula (1) can be transformed into:

[0049]

[0050] If X j is the value of each factor of the observation object in the non-exposure group, and X i is the value of each factor of the observation object in the exposure group, then the relative risk RR of the exposure group to the non-exposure group can be obtained from formula (3):

[0051]

[0052] Therefore:

[0053] When the covariate X j , takes 1 and 0 respectively, the corresponding RR j is:

[0054] RR j = exp(β j )

[0055] From formula (1) and formula (4), we can see that:

[0056] If β j > 0, RR j > 1, then the larger the value of each X j , the larger the value of h(t, X), that is, X jis a special factor;

[0057] If β j = 0, RR j = 1, the value of each X j has no effect on the value of h(t, X), that is, X j is an irrelevant factor;

[0058] If β j < 0, RR j < 1, the greater the value of each X j , the smaller the value of h(t, X), that is, X j is a basic factor.

[0059] Through the COX regression model, it can be concluded that the basic factors include accident type, vehicle attribute, accident level, vehicle brand, vehicle configuration, vehicle type, fuel type, surrounding temperature, weather, season and road condition information; the special factors include distance factor, driver population tendency industry, driver population tendency gender and driving behavior habit; the adjustment factors include rescue order receiving personnel good comment degree label, including order receiving rate, good comment rate and professionalism, and the rescue order receiving personnel good comment degree label is valued, such as a good comment rate of 90% or more is valued 5, and the specific value can be set according to the actual situation.

[0060] In this embodiment, the accident type includes anchor throwing, no electricity, emergency braking, tire, appearance damage, oil tank meter display, abnormal sound and central control operation; the accident level includes major, large, medium and small; the vehicle brand includes high-end car, medium-end car and low-end car; the vehicle configuration includes high-end car, medium-end car and low-end car; the vehicle type includes fuel car and electric car; the fuel type includes no fuel, 98# fuel, 95# fuel, 92# fuel and diesel; the weather includes extreme climate, ordinary rain and snow and no special; the season includes spring, summer, autumn and winter; the road condition information includes mountain, city, suburb, highway or expressway, country road, track and tunnel.

[0061] Furthermore, in this embodiment, the following are assigned values: breakdown, power failure, and emergency braking are assigned a value of 5; tires, fuel tank / electric meter display, and abnormal noise are assigned a value of 3; and exterior damage and central control operation are assigned a value of 1. For vehicle attributes, new cars (1-3 years old, no accidents) are assigned a value of 1; cars (3-7 years old, no accidents) are assigned a value of 2; cars (over 7 years old, no accidents) are assigned a value of 3; cars with minor accidents are assigned a value of 5; and cars with major accidents are assigned a value of 10. For accident levels, major accidents are assigned a value of 10; large accidents are assigned a value of 5; medium accidents are assigned a value of 3; and small accidents are assigned a value of 1. For vehicle brands, high-end cars are assigned a value of 5; mid-range cars are assigned a value of 3; and low-end cars are assigned a value of 1. For vehicle configurations, high-end cars are assigned a value of 5; mid-range cars are assigned a value of 3; and low-end cars are assigned a value of 1. For vehicle types, gasoline vehicles are assigned a value of 1, and electric vehicles are assigned a value of 1. The value is 5; for fuel types, no fuel is assigned a value of 5, 98-octane fuel is assigned a value of 4, 95-octane fuel is assigned a value of 3, 92-octane fuel is assigned a value of 2, and diesel is assigned a value of 1; for ambient temperatures below -20 degrees Celsius, a value of 5 is assigned, -10 to -20 degrees Celsius is assigned a value of 4; -10 to 0 degrees Celsius is assigned a value of 3, 0 to 10 degrees Celsius is assigned a value of 1, 10 to 30 degrees Celsius is assigned a value of 2, and above 30 degrees Celsius is assigned a value of 5; for weather, extreme weather is assigned a value of 5, ordinary rain and snow is assigned a value of 3, and no special weather is assigned a value of 1; for seasons, spring and autumn are assigned a value of 1, summer and winter are assigned a value of 2; for road condition information, mountainous areas are assigned a value of 3, urban areas are assigned a value of 1, suburbs, highways or expressways and rural roads are assigned a value of 2, and railways and tunnels are assigned a value of 5.

[0062] In this embodiment, the special factor is the information of the rescue order taker. A score of 10 is assigned when the distance between the rescue order taker and the user is less than 1km, 8 for 1km to 5km, 7 for 5km to 10km, 5 for more than 10km, and 1 for more than 30km. A score of 5 is assigned to rescue order takers who are aligned with the industry they belong to, and 1 for those who are not. A score of 5 is assigned to rescue order takers who are aligned with the gender they belong to, and 1 for those who are not. It is understood that scores can also be assigned to the rescue order taker's habits and operational behaviors based on actual circumstances.

[0063] Furthermore, when a user initiates a rescue, the system first analyzes the user's information and the rescue information. Then, based on the scores assigned by the aforementioned basic and special factors and the adjustment factors of rescue personnel in the rescue personnel tag library, the system uses a call matching algorithm to match the most suitable rescue personnel and accurately pushes the rescue request information.

[0064] In this embodiment, the formula for the call matching algorithm is:

[0065]

[0066] In the formula, SF is the matching score for rescue order takers, and a1, a2, ..., a n Basic factors, h1, h2, ..., hn b1, b2, …, b are the matching weights corresponding to the basic factors m k1, k2, …, k are the special factors m Y is the adjustment factor.

[0067] In the embodiment, the basic factor matching weight h n and the special factor matching weight k m are calculated according to the following formula:

[0068] Matching weight = the number of orders matched by the rescue order matching personnel corresponding to the factor / the total number of orders of the rescue order matching personnel.

[0069] In the embodiment, the top 5 (N=5) most suitable rescue order matching personnel are matched and locked by the call matching algorithm, and the rescue demand information is accurately distributed to the top 5 most suitable rescue order matching personnel.

[0070] In the embodiment, the order confirmation includes the following sub-steps:

[0071] If someone accepts the order within 10 (M=10) minutes, the rescue order matching is successful, if no one accepts the order, the rescue order matching personnel are diffused according to the calculation results of the call matching algorithm, and the top 10 most suitable rescue order matching personnel are diffused, and the top 20 most suitable rescue order matching personnel are diffused, etc. (It can be set according to the actual situation)

[0072] The rescue order matching personnel who receive the rescue demand information determine the rescue order matching personnel who successfully accept the order according to the order acceptance speed;

[0073] If an order in the breach state appears, according to real-time information, the most suitable rescue order matching personnel are matched in urgency by further diffusing the rescue order matching personnel, and the user is compensated.

[0074] In the embodiment, the rescue order matching personnel include official rescue personnel and ordinary users.

[0075] The application matches all the users (including official and ordinary users) of the voluntary system in the surrounding area through the self-created call matching algorithm SF, expands the order acceptance range of vehicle rescue demand, matches the most suitable vehicle rescue personnel, improves the vehicle rescue success rate, and thus improves the vehicle after-sales user satisfaction. The original single professional rescue team can be expanded to official rescue, ordinary users and other multi-dimensional rescue, and the rescue order acceptance rate is improved. The application realizes accurate matching based on the label system and big data technology, accurately pushes the content to the corresponding order acceptance user, improves the order acceptance quantity of the user, and reduces the order acceptance response time. The application expands the original business of merchants-users to the business of (merchants+users)-users, changes the overall logic of the original business, and also increases the social attribute.

[0076] It should be noted that, for the foregoing embodiments, for the sake of simple description, they are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited by the order of the described actions, because according to the present application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should know that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to the present application.

[0077] In the above embodiments, the basic principles and main features of the present application and the advantages of the present application are described. Those skilled in the art should understand that the present application is not limited by the above embodiments, and the above embodiments and the description in the specification are only to illustrate the principles of the present application. Any modification and change made by those skilled in the art without departing from the spirit and scope of the present application shall be within the protection scope of the claims of the present application.

Claims

1. A method for improving the success rate of vehicle rescue based on big data, characterized in that, Includes the following steps: Factor selection: All basic tags were obtained from the road rescue big data system tag library as enumeration factors; Factor classification: The enumerated factors are classified into three categories: basic factors, special factors, and moderating factors using the Cox regression model, and values ​​are assigned to the basic factors, special factors, and moderating factors respectively. Initiating a rescue request: The user fills in their user information and the user's rescue request details; Matching rescue personnel: Based on user information and the user's rescue information, the most suitable rescue personnel are matched through a call matching algorithm, and the rescue request information is accurately pushed to the user. Confirming Order Acceptance: After receiving the rescue request information, the rescue personnel will confirm whether to accept the order and then head to the accident site to carry out the rescue. The formula for the call matching algorithm is: ； In the formula, To assign points to rescue workers who accept orders, , , ..., Basic factors , , ..., To match weights to the corresponding basic factors, , , ..., As a special factor, , , ..., To match weights to specific factors, As a regulating factor; The basic factor matching weight Matching weights with special factors The calculation formulas are all: Matching weight = Number of orders already accepted by the rescue order taker according to the corresponding factor / Total number of orders of the rescue order taker.

2. The method for improving vehicle rescue success rate based on big data as described in claim 1, characterized in that, The formula for the Cox regression model is: ； In the formula, Let be the hazard rate function. It is when the X vector is 0. The baseline risk rate, , , ..., as independent variable The partial regression coefficient.

3. The method for improving vehicle rescue success rate based on big data as described in claim 1, characterized in that, The basic factors include accident type, vehicle attributes, accident level, vehicle brand, vehicle configuration, vehicle type, fuel type, ambient temperature, weather, season, and road condition information; the special factors include distance factors, driver demographics by industry, driver demographics by gender, and driving behavior habits; the adjustment factors include the rating tag of the rescue order taker.

4. The method for improving vehicle rescue success rate based on big data as described in claim 3, characterized in that, The accident types include breakdown, power failure, emergency braking, tire damage, exterior damage, fuel tank and battery meter readings, abnormal noises, and central control operation issues; the accident levels include major, large, medium, and small; the vehicle brands include high-end, mid-range, and low-end vehicles; the vehicle configurations include high-spec, mid-range, and low-spec vehicles; the vehicle types include gasoline vehicles and electric vehicles; the fuel types include no gasoline, 98-octane, 95-octane, 92-octane, and diesel; the weather includes extreme weather, ordinary rain and snow, and no special weather; the seasons include spring, summer, autumn, and winter; and the road condition information includes mountainous terrain, urban areas, suburbs, highways or expressways, rural roads, railways, and tunnels.

5. The method for improving vehicle rescue success rate based on big data as described in claim 1, characterized in that, The call matching algorithm identifies and identifies the top N most suitable rescue personnel, and the rescue request information is accurately sent to the top N most suitable rescue personnel.

6. The method for improving vehicle rescue success rate based on big data as described in claim 1, characterized in that, The order confirmation process includes the following sub-steps: If someone accepts the order within M minutes, the rescue order is successfully matched; if no one accepts the order, the rescue order is distributed to potential recipients based on the call matching algorithm. Once a rescue request is received, the rescue personnel who will successfully accept the order will be determined based on the speed at which the order is accepted. If an order is cancelled, the most suitable rescue personnel will be matched urgently based on real-time information, and the user will be compensated.

Images