Data analysis and commodity recommendation method and system based on machine learning

Abstract

The invention discloses a data analysis and commodity recommendation method and system based on machine learning, and belongs to the technical field of machine learning, and the method comprises the following steps: analyzing the IP matching degree, the equipment matching degree and the contact information matching degree between accounts; analyzing the comprehensive matching degree between the accounts; analyzing whether the accounts belong to the same user; collecting consumption records of all accounts belonging to the same user to form a consumption record set of the corresponding user; analyzing consumption characteristics of each user; the commodity consumption characteristics of all the users are analyzed, and a recommendation list is formed for commodity recommendation; according to the invention, through calculation of the comprehensive matching degree, a plurality of accounts belonging to the same user are identified, and a complete consumption record of the user is formed, so that portraits of the user are described more accurately, and commodity recommendation is more comprehensive and personalized; meanwhile, the computing resource management of the system is optimized, the computing power consumption of the system is reduced, and the stability and practicability of the system are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of machine learning, and specifically relates to a data analysis and product recommendation method and system based on machine learning. Background Art

[0002] With the rapid development of science and technology, the continuous upgrading of networks and smart devices has promoted the rapid rise and development of e-commerce. The number of users in the field of e-commerce has continued to increase, and the types of goods have become more and more diverse. With the popularization of e-commerce, it has become an indispensable part of people's lives.

[0003] E-commerce platforms mainly use users' historical shopping records to achieve user portraits and product recommendations. However, on e-commerce platforms, users often use multiple accounts for different purposes, such as obtaining discounts. These accounts appear to be independent accounts, but may actually belong to the same user. This leads to inaccurate portraits and incomplete product recommendations, while increasing the computing power consumption of the system and increasing the system's resource consumption.

[0004] Therefore, people urgently need data analysis and product recommendation systems based on machine learning to solve the above problems. Summary of the invention

[0005] The purpose of the present invention is to provide a data analysis and product recommendation system based on machine learning to solve the problems raised in the above background technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] The data analysis and product recommendation method based on machine learning includes the following steps:

[0008] S1. Collect login data of each account and analyze the IP matching, device matching and contact matching between each account;

[0009] S2. Analyze the comprehensive matching degree between the accounts based on the IP matching degree, device matching degree and contact matching degree between the accounts; and analyze whether the accounts belong to the same user based on the comprehensive matching degree;

[0010] S3. Collect consumption records of all accounts belonging to the same user to form a consumption record set of the corresponding user;

[0011] S4. Analyze the consumption characteristics of each user based on the consumption record set of each user;

[0012] S5. Analyze the commodity consumption characteristics of each user and form a recommendation list for commodity recommendations.

[0013] According to the above technical solution, the specific process of step S1 is as follows:

[0014] S1-1. Collect login data of each account, including IP address, login device identification and user contact information;

[0015] The login data is data information collected after obtaining user authorization;

[0016] The IP addresses of each account are counted to count the number of IP addresses logged in by the same account; the IP addresses of different accounts are screened to screen out the same IP addresses between different accounts as the IP addresses that appear in common between different accounts, and the number of IP addresses that appear in common between different accounts is recorded;

[0017] The number of IP addresses logged in by the ath account is recorded as A1 a ; Record the number of IP addresses logged in by the bth account as A2 b ; The number of IP addresses that appear in common between the ath account and the bth account is recorded as A3 a，b According to A1 a 、A2 b and A3 a，b , calculate the IP matching degree between the ath account and the bth account. The specific calculation formula is: A4 a，b =A3 a，b / (A1 a +A2 b ); A4 a，b Indicates the IP matching degree between the a-th account and the b-th account;

[0018] S1-2. Count the login device identification of each account and count the number of devices logged in by the same account; filter the login device identifications of different accounts and filter out devices with the same login device identification between different accounts as devices commonly used by different accounts, and record the number of devices commonly used by different accounts;

[0019] The number of devices logged in by the ath account is recorded as B1 a ; Record the number of devices logged in by the bth account as B2 b ; The number of devices used by account a and account b is recorded as B3 a，b According to B1 a 、B2 b and B3 a，b , calculate the device matching degree between the ath account and the bth account, the specific calculation formula is: B4 a，b =B3 a，b / (B1 a+B2 b ); among which, B4 a，b Indicates the device matching degree between the a-th account and the b-th account;

[0020] S1-3, compare the strings of user contact information of different accounts. If the strings corresponding to the user contact information of different accounts are the same, set the contact information matching degree between the corresponding accounts to α1; if the strings corresponding to the user contact information of different accounts are different, set the contact information matching degree between the corresponding accounts to α2; wherein α1 and α2 are system preset parameters;

[0021] The contact matching degree between the a-th account and the b-th account is recorded as C a，b ;

[0022] By analyzing the IP addresses, login device identification and user contact information between accounts, the similarities between accounts can be effectively obtained, which effectively improves the accuracy of subsequent analysis.

[0023] According to the above technical solution, the specific process of step S2 is as follows:

[0024] S2-1, based on the IP matching degree A4 between the ath account and the bth account a，b , the device matching degree between the ath account and the bth account B4 a，b And the contact matching degree C between the ath account and the bth account a，b , calculate the comprehensive matching degree between the ath account and the bth account, the specific calculation formula is: D a，b =β1×A4 a，b +β2×B4 a，b +β3×C a，b ; Among them, D a，b represents the comprehensive matching degree between the a-th account and the b-th account; β1 represents the weight coefficient of IP matching degree; β2 represents the weight coefficient of device matching degree; β3 represents the weight coefficient of contact information matching degree;

[0025] S2-2. If the comprehensive matching degree between the a-th account and the b-th account is D a，b If it is greater than or equal to γ, it means that the a-th account and the b-th account belong to the same user; where γ represents a system preset parameter;

[0026] If the comprehensive matching degree between the ath account and the bth account is D a，b If it is less than γ, it means that the a-th account and the b-th account do not belong to the same user;

[0027] The comprehensive matching degree of the accounts is calculated by comprehensively considering the IP matching degree, device matching degree and contact matching degree. Based on the comprehensive matching degree, it is determined whether each account belongs to the same user, thereby reducing the misjudgment rate.

[0028] According to the above technical solution, the specific process of step S3 is as follows:

[0029] S3-1. The consumption record includes consumption time, consumption merchant and commodity type;

[0030] S3-2, collecting consumption records of all accounts belonging to the same user, and arranging the consumption records of all accounts belonging to the same user according to the order of consumption time to form a consumption record set of the corresponding user;

[0031] By using all accounts belonging to the same user, we can build up the user’s complete consumption history and provide basic data for consumption feature analysis, making the user portrait more accurate and product recommendations more comprehensive.

[0032] According to the above technical solution, the specific process of step S4 is as follows:

[0033] S4-1. The consumption characteristics include merchant recognition, product preference and discount preference;

[0034] S4-2. The consumption record set of the i-th user is recorded as set E i ; According to the set E i , count the number of times the ith user spends in different consumer merchants, and the total number of times the ith user spends in all consumer merchants; calculate the ratio of the number of times the ith user spends in different consumer merchants to the total number of times the ith user spends in all consumer merchants, as the merchant recognition of the ith user for different consumer merchants;

[0035] S4-3. According to the set E i , count the types of goods purchased by the ith user in different consumer merchants; according to the types of goods, count the number of times the ith user purchases different types of goods and the total number of times the ith user purchases all types of goods; calculate the ratio of the number of times the ith user purchases different types of goods to the total number of times the ith user purchases all types of goods as the ith user's preference for different types of goods;

[0036] S4-4. According to the set E i, count the consumption time of the i-th user in different consumer merchants; collect the discount time of all consumer merchants corresponding to the i-th user, combine the consumption time of the i-th user in different consumer merchants, match the consumption records of the i-th user in different consumer merchants whose consumption time is within the discount time of the corresponding merchant, and record the number of consumption records matched by the i-th user; calculate the number of consumption records matched by the i-th user and the set E i The ratio of the total number of consumption records in is taken as the discount preference of the i-th user;

[0037] Analyze users' consumption characteristics, including merchant recognition, product preference and discount preference, to gain an in-depth understanding of users' consumption habits and preferences, and lay the foundation for subsequent product recommendations.

[0038] According to the above technical solution, the specific process of step S5 is as follows:

[0039] S5-1. The merchant recognition of the i-th user for the j-th consumer merchant is recorded as F1 i，j ; The liking degree of the i-th user for the k-th product in the j-th consumer merchant is recorded as F2 i，j，k ;

[0040] Calculate the recommendation degree of the kth product in the jth consumer merchant for the ith user. The specific calculation formula is: F3 i，j，k =δ1×F1 i，j +δ2×F2 i，j，k Among them, F3 i，j，k represents the recommendation degree of the kth product in the jth consumer merchant to the ith user; δ1 represents the weight coefficient of recognition; δ2 represents the weight coefficient of liking;

[0041] S5-2. If the discount preference of the i-th user is less than ε, then according to the recommendation degree F3 of the k-th product in the j-th consumer merchant for the i-th user i，j，k Arrange the products from high to low to form a recommendation list, and send the recommendation list to all accounts belonging to the i-th user at the same time to make product recommendations; where ε represents the system preset parameter;

[0042] S5-3. If the discount preference of the i-th user is greater than or equal to ε, the discount ratio of the k-th commodity in the j-th consumer merchant is collected, denoted as G j，k According to F3 i，j，k and G2 j，k , calculate the adjusted recommendation degree of the kth product in the jth consumer merchant for the ith user. The specific calculation formula is: F4 i，j，k =F3 i，j，k +δ3×(1-G2 j，k ), where F4i，j，k represents the adjusted recommendation degree of the kth product in the jth consumer merchant for the ith user; δ3 represents the weight coefficient of the discount ratio;

[0043] According to the k-th product in the j-th consumer merchant, the adjusted recommendation degree F4 for the i-th user i，j，k Arrange the products from high to low to form a recommendation list, and send the recommendation list to all accounts belonging to the i-th user at the same time to make product recommendations;

[0044] Generate a recommendation list based on the user's consumption characteristics and discount preferences to improve the user experience. At the same time, make product recommendations for all accounts belonging to the same user, reduce the system's computing power consumption, and reduce the system's resource consumption.

[0045] A data analysis and product recommendation system based on machine learning, the system includes an information collection module, an information processing module and an execution module;

[0046] The information acquisition module is used to acquire the information required by the system; the information processing module is used to store, analyze and transmit the information of each module; and the execution module is used to execute the information of the information processing module.

[0047] According to the above technical solution, the information collection module includes an account information unit and a merchant information unit;

[0048] The account information unit is used to collect account login data and consumption records; the merchant information unit is used for merchant data information.

[0049] According to the above technical solution, the information processing module includes an information storage unit, an information analysis unit and an information transmission unit;

[0050] The information storage unit is used to store the information acquired by the information acquisition module; the information analysis unit is used to analyze the information acquired by the information acquisition module; and the information transmission unit is used for information transmission of each module in the system.

[0051] According to the above technical solution, the execution module includes a display unit and a control unit;

[0052] The display unit is used to display a recommendation list of commodities to form visual data; the control unit is used to control the sending of the recommendation list.

[0053] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0054] The present invention realizes accurate identification of accounts by matching and analyzing data such as the account's IP address, login device identifier, and user contact information. At the same time, the system intelligently identifies multiple accounts belonging to the same user by calculating the comprehensive matching degree, thereby avoiding data redundancy and repeated analysis, and improving the efficiency and accuracy of data processing; through all accounts belonging to the same user, the user's complete consumption history is constructed to ensure the comprehensiveness of data analysis, making the user's portrait more accurate and the product recommendation more comprehensive; in addition, the present invention realizes personalized product recommendation through in-depth analysis of the user's consumption characteristics, and can improve the accuracy and conversion rate of recommendations under the user's specific preferences, thereby improving the user experience; while improving the accuracy of user portrait portrayal and the comprehensiveness of product recommendations, the present invention optimizes the system's computing resource management, reduces the system's computing power consumption, and improves the system's stability and practicality. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0056] Figure 1 It is a flow chart of the data analysis and product recommendation method based on machine learning of the present invention;

[0057] Figure 2 It is a structural diagram of the data analysis and product recommendation system based on machine learning of the present invention. DETAILED DESCRIPTION

[0058] 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.

[0059] See also Figure 1 , the present invention provides a technical solution:

[0060] The data analysis and product recommendation method based on machine learning includes the following steps:

[0061] S1. Collect login data of each account and analyze the IP matching, device matching and contact matching between each account;

[0062] S2. Analyze the comprehensive matching degree between the accounts based on the IP matching degree, device matching degree and contact matching degree between the accounts; and analyze whether the accounts belong to the same user based on the comprehensive matching degree;

[0063] S3. Collect consumption records of all accounts belonging to the same user to form a consumption record set of the corresponding user;

[0064] S4. Analyze the consumption characteristics of each user based on the consumption record set of each user;

[0065] S5. Analyze the commodity consumption characteristics of each user and form a recommendation list for commodity recommendations.

[0066] According to the above technical solution, the specific process of step S1 is as follows:

[0067] S1-1. Collect login data of each account, including IP address, login device identification and user contact information;

[0068] The login data is data information collected after obtaining user authorization;

[0069] The IP addresses of each account are counted to count the number of IP addresses logged in by the same account; the IP addresses of different accounts are screened to screen out the same IP addresses between different accounts as the IP addresses that appear in common between different accounts, and the number of IP addresses that appear in common between different accounts is recorded;

[0070] The number of IP addresses logged in by the ath account is recorded as A1 a ; Record the number of IP addresses logged in by the bth account as A2 b ; The number of IP addresses that appear in common between the ath account and the bth account is recorded as A3 a，b According to A1 a 、A2 b and A3 a，b , calculate the IP matching degree between the ath account and the bth account. The specific calculation formula is: A4 a，b =A3 a，b / (A1 a +A2 b ); A4 a，b Indicates the IP matching degree between the a-th account and the b-th account;

[0071] S1-2. Count the login device identification of each account and count the number of devices logged in by the same account; filter the login device identifications of different accounts and filter out devices with the same login device identification between different accounts as devices commonly used by different accounts, and record the number of devices commonly used by different accounts;

[0072] The number of devices logged in by the ath account is recorded as B1 a; Record the number of devices logged in by the bth account as B2 b ; The number of devices used by account a and account b is recorded as B3 a，b According to B1 a 、B2 b and B3 a，b , calculate the device matching degree between the ath account and the bth account, the specific calculation formula is: B4 a，b =B3 a，b / (B1 a +B2 b ); among which, B4 a，b Indicates the device matching degree between the a-th account and the b-th account;

[0073] S1-3, compare the strings of user contact information of different accounts. If the strings corresponding to the user contact information of different accounts are the same, set the contact information matching degree between the corresponding accounts to α1; if the strings corresponding to the user contact information of different accounts are different, set the contact information matching degree between the corresponding accounts to α2; wherein α1 and α2 are system preset parameters;

[0074] The contact matching degree between the a-th account and the b-th account is recorded as C a，b ;

[0075] By analyzing the IP addresses, login device identification and user contact information between accounts, the similarities between accounts can be effectively obtained, which effectively improves the accuracy of subsequent analysis.

[0076] According to the above technical solution, the specific process of step S2 is as follows:

[0077] S2-1, based on the IP matching degree A4 between the ath account and the bth account a，b , the device matching degree between the ath account and the bth account B4 a，b And the contact matching degree C between the ath account and the bth account a，b , calculate the comprehensive matching degree between the ath account and the bth account, the specific calculation formula is: D a，b =β1×A4 a，b +β2×B4 a，b +β3×C a，b ; Among them, D a，b represents the comprehensive matching degree between the a-th account and the b-th account; β1 represents the weight coefficient of IP matching degree; β2 represents the weight coefficient of device matching degree; β3 represents the weight coefficient of contact information matching degree;

[0078] S2-2. If the comprehensive matching degree between the a-th account and the b-th account is Da，b If it is greater than or equal to γ, it means that the a-th account and the b-th account belong to the same user; where γ represents a system preset parameter;

[0079] If the comprehensive matching degree between the ath account and the bth account is D a，b If it is less than γ, it means that the a-th account and the b-th account do not belong to the same user;

[0080] For example:

[0081] The number of IP addresses logged in by the ath account A1 a =3; the number of IP addresses logged in by the bth account A2 b =2; the number of IP addresses that appear in common between the ath account and the bth account A3 a，b =2;

[0082] According to A1 a 、A2 b and A3 a，b , calculate the IP matching degree between the ath account and the bth account. The specific calculation formula is: A4 a，b =A3 a，b / (A1 a +A2 b )=2 / (3+2)=0.4;

[0083] The number of devices logged in by the ath account is recorded as B1 a =2; the number of devices logged in by the bth account B2 b =2; the number of devices used by the ath account and the bth account in common is B3 a，b =1; According to B1 a 、B2 b and B3 a，b , calculate the device matching degree between the ath account and the bth account, the specific calculation formula is: B4 a，b =B3 a，b / (B1 a +B2 b ) = 1 / (2+2) = 0.25;

[0084] Compare the strings of the user contact information of the ath account and the bth account. If the strings corresponding to the user contact information of the ath account and the bth account are different, the matching degree of the contact information between the corresponding accounts is set to α1=0;

[0085] It is known that β1=0.4, β2=0.4, β3=0.2, γ=0.5;

[0086] Calculate the comprehensive matching degree between the ath account and the bth account. The specific calculation formula is: Da，b =β1×A4 a，b +β2×B4 a，b +β3×C a，b =0.4×0.4+0.4×0.25+0.2×0=0.26;

[0087] The comprehensive matching degree D between the ath account and the bth account a，b =0.26 is less than γ=0.5, indicating that the a-th account and the b-th account do not belong to the same user.

[0088] The comprehensive matching degree of the accounts is calculated by comprehensively considering the IP matching degree, device matching degree and contact matching degree. Based on the comprehensive matching degree, it is determined whether each account belongs to the same user, thereby reducing the misjudgment rate.

[0089] According to the above technical solution, the specific process of step S3 is as follows:

[0090] S3-1. The consumption record includes consumption time, consumption merchant and commodity type;

[0091] S3-2, collecting consumption records of all accounts belonging to the same user, and arranging the consumption records of all accounts belonging to the same user according to the order of consumption time to form a consumption record set of the corresponding user;

[0092] By using all accounts belonging to the same user, we can build up the user’s complete consumption history and provide basic data for consumption feature analysis, making the user portrait more accurate and product recommendations more comprehensive.

[0093] According to the above technical solution, the specific process of step S4 is as follows:

[0094] S4-1. The consumption characteristics include merchant recognition, product preference and discount preference;

[0095] S4-2. The consumption record set of the i-th user is recorded as set E i ; According to the set E i , count the number of times the ith user spends in different consumer merchants, and the total number of times the ith user spends in all consumer merchants; calculate the ratio of the number of times the ith user spends in different consumer merchants to the total number of times the ith user spends in all consumer merchants, as the merchant recognition of the ith user for different consumer merchants;

[0096] S4-3. According to the set E i, count the types of goods purchased by the ith user in different consumer merchants; according to the types of goods, count the number of times the ith user purchases different types of goods and the total number of times the ith user purchases all types of goods; calculate the ratio of the number of times the ith user purchases different types of goods to the total number of times the ith user purchases all types of goods as the ith user's preference for different types of goods;

[0097] S4-4. According to the set E i , count the consumption time of the i-th user in different consumer merchants; collect the discount time of all consumer merchants corresponding to the i-th user, combine the consumption time of the i-th user in different consumer merchants, match the consumption records of the i-th user in different consumer merchants whose consumption time is within the discount time of the corresponding merchant, and record the number of consumption records matched by the i-th user; calculate the number of consumption records matched by the i-th user and the set E i The ratio of the total number of consumption records in is taken as the discount preference of the i-th user;

[0098] Analyze users' consumption characteristics, including merchant recognition, product preference and discount preference, to gain an in-depth understanding of users' consumption habits and preferences, and lay the foundation for subsequent product recommendations.

[0099] According to the above technical solution, the specific process of step S5 is as follows:

[0100] S5-1. The merchant recognition of the i-th user for the j-th consumer merchant is recorded as F1 i，j ; The liking degree of the i-th user for the k-th product in the j-th consumer merchant is recorded as F2 i，j，k ;

[0101] Calculate the recommendation degree of the kth product in the jth consumer merchant for the ith user. The specific calculation formula is: F3 i，j，k =δ1×F1 i，j +δ2×F2 i，j，k Among them, F3 i，j，k represents the recommendation degree of the kth product in the jth consumer merchant to the ith user; δ1 represents the weight coefficient of recognition; δ2 represents the weight coefficient of liking;

[0102] S5-2. If the discount preference of the i-th user is less than ε, then according to the recommendation degree F3 of the k-th product in the j-th consumer merchant for the i-th user i，j，k Arrange the products from high to low to form a recommendation list, and send the recommendation list to all accounts belonging to the i-th user at the same time to make product recommendations; where ε represents the system preset parameter;

[0103] S5-3. If the discount preference of the i-th user is greater than or equal to ε, the discount ratio of the k-th commodity in the j-th consumer merchant is collected, denoted as G j，k According to F3 i，j，k and G2 j，k , calculate the adjusted recommendation degree of the kth product in the jth consumer merchant for the ith user. The specific calculation formula is: F4 i，j，k =F3 i，j，k +δ3×(1-G2 j，k ), where F4 i，j，k represents the adjusted recommendation degree of the kth product in the jth consumer merchant for the ith user; δ3 represents the weight coefficient of the discount ratio;

[0104] According to the k-th product in the j-th consumer merchant, the adjusted recommendation degree F4 for the i-th user i，j，k Arrange the products from high to low to form a recommendation list, and send the recommendation list to all accounts belonging to the i-th user at the same time to make product recommendations;

[0105] Generate a recommendation list based on the user's consumption characteristics and discount preferences to improve the user experience. At the same time, make product recommendations for all accounts belonging to the same user, reduce the system's computing power consumption, and reduce the system's resource consumption.

[0106] See also Figure 2 , a data analysis and product recommendation system based on machine learning, the system includes an information collection module, an information processing module and an execution module;

[0107] The information acquisition module is used to acquire the information required by the system; the information processing module is used to store, analyze and transmit the information of each module; and the execution module is used to execute the information of the information processing module.

[0108] According to the above technical solution, the information collection module includes an account information unit and a merchant information unit;

[0109] The account information unit is used to collect account login data and consumption records; the merchant information unit is used for merchant data information.

[0110] According to the above technical solution, the information processing module includes an information storage unit, an information analysis unit and an information transmission unit;

[0111] The information storage unit is used to store the information acquired by the information acquisition module; the information analysis unit is used to analyze the information acquired by the information acquisition module; and the information transmission unit is used for information transmission of each module in the system.

[0112] According to the above technical solution, the execution module includes a display unit and a control unit;

[0113] The display unit is used to display a recommendation list of commodities to form visual data; the control unit is used to control the sending of the recommendation list.

[0114] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions recorded in the aforementioned embodiments, or to replace some of the technical features therein by equivalents. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A data analysis and product recommendation method based on machine learning, characterized in that: The method comprises the following steps: S1. Collect login data of each account and analyze the IP matching, device matching and contact matching between each account; S2. Analyze the comprehensive matching degree between the accounts based on the IP matching degree, device matching degree and contact matching degree between the accounts; and analyze whether the accounts belong to the same user based on the comprehensive matching degree between the accounts; S3. Collect consumption records of all accounts belonging to the same user to form a consumption record set of the corresponding user; S4. Analyze the consumption characteristics of each user based on the consumption record set of each user; S5. Analyze the commodity consumption characteristics of each user and form a recommendation list for commodity recommendations.

2. The data analysis and product recommendation method based on machine learning according to claim 1, characterized in that: The specific process of step S1 is as follows: S1-1. Collect login data of each account, including IP address, login device identification and user contact information; Collect statistics on the IP addresses of each account and count the number of IP addresses logged in by the same account; Filter the IP addresses of different accounts, filter out the same IP addresses between different accounts, and use them as the IP addresses that appear in common between different accounts. Record the number of IP addresses that appear in common between different accounts. The number of IP addresses logged in by the ath account is recorded as A1 a ; Record the number of IP addresses logged in by the bth account as A2 b ; The number of IP addresses that appear in common between the ath account and the bth account is recorded as A3 a，b According to A1 a 、A2 b and A3 a，b , calculate the IP matching degree between the ath account and the bth account. The specific calculation formula is: A4 a，b =A3 a，b / (A1 a +A2 b ); A4 a，b Indicates the IP matching degree between the a-th account and the b-th account; S1-2. Count the login device identification of each account and count the number of devices logged in by the same account; filter the login device identifications of different accounts and filter out devices with the same login device identification between different accounts as devices commonly used by different accounts, and record the number of devices commonly used by different accounts; The number of devices logged in by the ath account is recorded as B1 a ; Record the number of devices logged in by the bth account as B2 b ; The number of devices used by account a and account b is recorded as B3 a，b According to B1 a 、B2 b and B3 a，b , calculate the device matching degree between the ath account and the bth account, the specific calculation formula is: B4 a，b =B3 a，b / (B1 a +B2 b ); among which, B4 a，b Indicates the device matching degree between the a-th account and the b-th account; S1-3, compare the strings of user contact information of different accounts. If the strings corresponding to the user contact information of different accounts are the same, set the contact information matching degree between the corresponding accounts to α1; if the strings corresponding to the user contact information of different accounts are different, set the contact information matching degree between the corresponding accounts to α2; wherein α1 and α2 are system preset parameters; The contact matching degree between the a-th account and the b-th account is recorded as C a，b .

3. The method for data analysis and product recommendation based on machine learning according to claim 2, characterized in that: The specific process of step S2 is as follows: S2-1, based on the IP matching degree A4 between the ath account and the bth account a，b , the device matching degree between the ath account and the bth account B4 a，b And the contact matching degree C between the ath account and the bth account a，b , calculate the comprehensive matching degree between the ath account and the bth account, the specific calculation formula is: D a，b =β1×A4 a，b +β2×B4 a，b +β3×C a，b ; Among them, D a，b represents the comprehensive matching degree between the a-th account and the b-th account; β1 represents the weight coefficient of IP matching degree; β2 represents the weight coefficient of device matching degree; β3 represents the weight coefficient of contact information matching degree; S2-2. If the comprehensive matching degree between the a-th account and the b-th account is D a，b If it is greater than or equal to γ, it means that the a-th account and the b-th account belong to the same user; where γ represents a system preset parameter; If the comprehensive matching degree between the ath account and the bth account is D a，b If it is less than γ, it means that the a-th account and the b-th account do not belong to the same user.

4. The method for data analysis and product recommendation based on machine learning according to claim 3, characterized in that: The specific process of step S3 is as follows: S3-1. The consumption record includes consumption time, consumption merchant and commodity type; S3-2. Collect consumption records of all accounts belonging to the same user, and arrange the consumption records of all accounts belonging to the same user according to the order of consumption time to form a consumption record set of the corresponding user.

5. The method for data analysis and product recommendation based on machine learning according to claim 4, characterized in that: The specific process of step S4 is as follows: S4-1. The consumption characteristics include merchant recognition, product preference and discount preference; S4-2. The consumption record set of the i-th user is recorded as set E i ; According to the set E i , count the number of times the ith user spends in different consumer merchants, and the total number of times the ith user spends in all consumer merchants; calculate the ratio of the number of times the ith user spends in different consumer merchants to the total number of times the ith user spends in all consumer merchants, as the merchant recognition of the ith user for different consumer merchants; S4-3. According to the set E i , count the types of goods purchased by the i-th user in different consumer merchants; According to the product types, count the number of times the i-th user purchases different types of products, as well as the total number of times the i-th user purchases all types of products; Calculate the ratio of the number of times the ith user purchases different types of goods to the total number of times the ith user purchases all types of goods, as the ith user's preference for different types of goods; S4-4. According to the set E i , count the consumption time of the i-th user in different consumer merchants; collect the discount time of all consumer merchants corresponding to the i-th user, combine the consumption time of the i-th user in different consumer merchants, match the consumption records of the i-th user in different consumer merchants whose consumption time is within the discount time of the corresponding merchant, and record the number of consumption records matched by the i-th user; Calculate the number of consumption records matched by the i-th user and the set E i The ratio of the total number of consumption records in is taken as the discount preference of the i-th user.

6. The method for data analysis and product recommendation based on machine learning according to claim 5, characterized in that: The specific process of step S5 is as follows: S5-1. The merchant recognition of the i-th user for the j-th consumer merchant is recorded as F1 i，j ; The liking degree of the i-th user for the k-th product in the j-th consumer merchant is recorded as F2 i，j，k ; Calculate the recommendation degree of the kth product in the jth consumer merchant for the ith user. The specific calculation formula is: F3 i，j，k =δ1×F1 i，j +δ2×F2 i，j，k Among them, F3 i，j，k represents the recommendation degree of the kth product in the jth consumer merchant to the ith user; δ1 represents the weight coefficient of recognition; δ2 represents the weight coefficient of liking; S5-2. If the discount preference of the i-th user is less than ε, then according to the recommendation degree F3 of the k-th product in the j-th consumer merchant for the i-th user i，j，k Arrange the products from high to low to form a recommendation list, and send the recommendation list to all accounts belonging to the i-th user at the same time to make product recommendations; where ε represents the system preset parameter; S5-3. If the discount preference of the i-th user is greater than or equal to ε, the discount ratio of the k-th commodity in the j-th consumer merchant is collected, denoted as G j，k According to F3 i，j，k and G2 j，k , calculate the adjusted recommendation degree of the kth product in the jth consumer merchant for the ith user. The specific calculation formula is: F4 i，j，k =F3 i，j，k +δ3×(1-G2 j，k ), where F4 i，j，k represents the adjusted recommendation degree of the kth product in the jth consumer merchant for the ith user; δ3 represents the weight coefficient of the discount ratio; According to the k-th product in the j-th consumer merchant, the adjusted recommendation degree F4 for the i-th user i，j，k The products are arranged from high to low to form a recommendation list, and the recommendation list is sent to all accounts belonging to the i-th user at the same time to make product recommendations.

7. A data analysis and product recommendation system based on machine learning, such as using the data analysis and product recommendation method based on machine learning according to any one of claims 1 to 6, characterized in that: The system includes an information collection module, an information processing module and an execution module; The information acquisition module is used to acquire the information required by the system; the information processing module is used to store, analyze and transmit the information of each module; and the execution module is used to execute the information of the information processing module.

8. The data analysis and product recommendation system based on machine learning according to claim 7, characterized in that: The information collection module includes an account information unit and a merchant information unit; The account information unit is used to collect account login data and consumption records; the merchant information unit is used for merchant data information.

9. The data analysis and product recommendation system based on machine learning according to claim 7, characterized in that: The information processing module includes an information storage unit, an information analysis unit and an information transmission unit; The information storage unit is used to store the information acquired by the information acquisition module; the information analysis unit is used to analyze the information acquired by the information acquisition module; and the information transmission unit is used for information transmission of each module in the system.

10. The data analysis and product recommendation system based on machine learning according to claim 7, characterized in that: The execution module includes a display unit and a control unit; The display unit is used to display a recommendation list of commodities to form visual data; the control unit is used to control the sending of the recommendation list.

Images