Marketing full-link analysis method and device based on artificial intelligence, equipment and medium

Abstract

The invention provides a marketing full-link analysis method, device, equipment and medium based on artificial intelligence, and the method comprises the steps: carrying out the automatic processing and analysis of a large amount of sales information through an analysis model based on time series data, and carrying out the normalization processing and threshold comparison, and carrying out the recognition of the sales information when an indication parameter is abnormal. A new simulation product vector is input into the model, parameters are adjusted, a reference product vector is compared with a target simulation product vector, and the marketing strength of various products is set according to a comparison result. The method has the beneficial effects that a commercial tenant can formulate a personalized marketing strategy for each type of products, and the pertinence and effectiveness of marketing are improved. By accurately analyzing consumer demands and market trends and formulating personalized marketing strategies, merchants can better meet the consumer demands, and brand images and market competitiveness are improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and in particular to an artificial intelligence-based marketing full-link analysis method, device, equipment and medium. Background Art

[0002] With the rapid development of artificial intelligence technology, its application in the field of marketing is becoming more and more extensive and in-depth, which is profoundly changing marketing strategies, consumer experience and business value. As an emerging marketing method, artificial intelligence marketing aims to improve marketing effectiveness and efficiency by using artificial intelligence technology to intelligently design and implement marketing activities.

[0003] At present, marketing is still at the stage of data analysis. It is impossible to achieve intelligent marketing by obtaining indicator data and then manually adjusting marketing strategies. Summary of the invention

[0004] The main purpose of the present invention is to provide a marketing full-link analysis method, device, equipment and medium based on artificial intelligence, aiming to solve the problem of being unable to achieve intelligent marketing.

[0005] The present invention provides a marketing full-link analysis method based on artificial intelligence, comprising:

[0006] Obtaining comment information from a preset comment database through a preset crawler network;

[0007] Setting target sales information for various products based on the review information;

[0008] Obtain sales information of various products at multiple time points within the time period set by the merchant;

[0009] Vectorize the sales information of various products at each time point to obtain the product vector corresponding to each time point;

[0010] Inputting each of the product vectors into a preset analysis model based on time series data in chronological order, and obtaining a sales status vector of the merchant within the set time period;

[0011] Normalizing the sales state vector to obtain corresponding indication parameters;

[0012] Compare the indication parameter with a set threshold; wherein, when the linear distance is higher than the set threshold, the indication parameter is determined to be abnormal, and the set threshold is obtained based on the sales information of various products at the first time point within the time period set by the merchant and the target sales information of various products;

[0013] When the indication parameter is abnormal, a new simulated product vector is input into the preset analysis model based on time series data, and the parameters of the simulated product vector are continuously adjusted until the obtained indication parameter is displayed as normal, thereby obtaining a target simulated product vector;

[0014] Extracting the product vector with the latest timing within the set time period as a reference product vector;

[0015] The reference product vector is compared with the target simulated product vector, and the marketing intensity of each product is set according to the comparison result.

[0016] Furthermore, the preset analysis model based on time series data is a long short-term memory network model, and the step of inputting each of the product vectors into the preset analysis model based on time series data in chronological order to obtain the sales state vector of the merchant within the set time period includes:

[0017] Input each of the product vectors into the forget gate, input gate, new information, and output gate of each recurrent unit in the long short-term memory network in sequence;

[0018] According to formula F t =sigmoid(W xf |X t |+W hf H t-1 +b f )

[0019] I t =sigmoid(W xi |X t |+W hi H t-1 +b i )

[0020] O t =sigmoid(W xo |X t |+W ho H t-1 +b o )

[0021] N t =tanh(W xn |X t |+W hn H t-1 +b n )

[0022] C t =F t ⊙C t-1 +I t ⊙Nt

[0023] Calculate the output state of each cycle unit in turn to obtain the sales state vector output by the last cycle unit; where F t represents the tth forget gate, I t represents the tth input gate, N t represents the tth new information, O t represents the tth output gate, C t represents the tth output state, X t represents the t-th product vector, W xf , W hf 、b f , W xi , W hi 、b i , W xo , W ho 、b o , W xn , W hn 、b n These are the preset parameters.

[0024] Furthermore, when the indication parameter is abnormal, a new simulated product vector is input into the preset analysis model based on time series data, and the parameters of the simulated product vector are continuously adjusted until the obtained indication parameter is displayed as normal, and the step of obtaining the target simulated product vector includes:

[0025] When the indication parameter is abnormal, one or more recurrent units are added to the long short-term memory network, and the simulated product vector is input into the newly added recurrent unit and the simulated product vector is continuously adjusted until the indication parameter is normal, thereby obtaining the final adjusted target simulated product vector.

[0026] Furthermore, the step of obtaining comment information from a preset comment database through a preset crawler network includes:

[0027] Obtaining location information of the preset comment database;

[0028] The corresponding comment information is obtained at the position corresponding to the position information through the sqoop script.

[0029] Furthermore, the step of setting target sales information of various products according to the review information includes:

[0030] Process each of the review information according to a preset natural language processing model and sentiment analysis model to analyze the customer's purchase expectations;

[0031] Target sales information for each product category is set based on historical sales information and the purchase expectations.

[0032] Furthermore, after the step of comparing the indication parameter with the set threshold, the method further includes:

[0033] When the indicated parameters are normal, the previous marketing plan is continued.

[0034] Furthermore, the step of comparing the reference product vector with the target simulated product vector and setting the marketing strength of each product according to the comparison result includes:

[0035] Calculating a difference vector between the reference product vector and the target simulated product vector;

[0036] Extracting first and second values ​​of various types of products from the difference vector and the reference product vector;

[0037] Compare the first value with the second value to obtain the ratio of each type of product;

[0038] The marketing intensity of each product category is set according to the ratio.

[0039] The present invention provides a marketing full-link analysis device based on artificial intelligence, comprising:

[0040] A first acquisition module is used to acquire comment information from a preset comment database through a preset crawler network;

[0041] A setting module, used to set target sales information of various products according to the review information;

[0042] The second acquisition module is used to obtain sales information of various products at multiple time points within a time period set by the merchant;

[0043] The vectorization module is used to vectorize the sales information of various products at various time points to obtain the product vectors corresponding to each time point;

[0044] A first input module, used to input each of the product vectors into a preset analysis model based on time series data in chronological order, to obtain a sales status vector of the merchant within the set time period;

[0045] A normalization module, used for normalizing the sales state vector to obtain corresponding indication parameters;

[0046] A first comparison module is used to compare the indication parameter with a set threshold value; wherein, when the linear distance is higher than the set threshold value, the indication parameter is determined to be abnormal, and the set threshold value is obtained based on the sales information of various products at the first time point within the time period set by the merchant and the target sales information of various products;

[0047] A second input module is used for inputting a new simulated product vector into the preset analysis model based on time series data when the indication parameter is abnormal, and continuously adjusting the parameters of the simulated product vector until the obtained indication parameter is displayed as normal, thereby obtaining a target simulated product vector;

[0048] An extraction module, used for extracting the product vector with the latest time sequence within the set time period as a reference product vector;

[0049] The second comparison module is used to compare the reference product vector with the target simulated product vector, and set the marketing intensity of each product according to the comparison result.

[0050] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.

[0051] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned methods are implemented.

[0052] The beneficial effects of the present invention are as follows: With the help of an analysis model based on time series data, a large amount of sales data can be automatically processed and analyzed, reducing manual intervention and improving analysis efficiency. By normalization processing and threshold comparison, when indicating abnormal parameters, by inputting new simulated product vectors into the model and adjusting parameters, by comparing the reference product vector with the target simulated product vector, merchants can formulate personalized marketing strategies for each type of product to improve the pertinence and effectiveness of marketing. By accurately analyzing consumer demand and market trends, and formulating personalized marketing strategies, merchants can better meet consumer demand and enhance brand image and market competitiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a flowchart of a marketing full-link analysis method based on artificial intelligence according to an embodiment of the present invention;

[0054] Figure 2 It is a schematic block diagram of the structure of a marketing full-link analysis device based on artificial intelligence according to an embodiment of the present invention;

[0055] Figure 3 A schematic block diagram of the structure of a computer device according to an embodiment of the present application.

[0056] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0057] 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.

[0058] It should be noted that all directional indications in the embodiments of the present invention (such as up, down, left, right, front, back, etc.) are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly. The connection can be a direct connection or an indirect connection.

[0059] The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there may be three relationships. For example, A and B can represent three situations: A exists alone, A and B exist at the same time, and B exists alone.

[0060] In addition, in the present invention, descriptions such as "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0061] Reference Figure 1 The present invention proposes a marketing full-link analysis method based on artificial intelligence, comprising:

[0062] S1: Obtain comment information from a preset comment database through a preset crawler network;

[0063] S2: setting target sales information for various products according to the review information;

[0064] S3: Obtain sales information of various products at multiple time points within the time period set by the merchant;

[0065] S4: vectorize the sales information of various products at each time point to obtain the product vector corresponding to each time point;

[0066] S5: Inputting each of the product vectors into a preset analysis model based on time series data in chronological order, and obtaining a sales status vector of the merchant within the set time period;

[0067] S6: normalizing the sales state vector to obtain corresponding indication parameters;

[0068] S7: comparing the indication parameter with a set threshold; wherein, when the linear distance is higher than the set threshold, the indication parameter is determined to be abnormal, and the set threshold is obtained based on the sales information of various products at the first time point within the time period set by the merchant and the target sales information of various products;

[0069] S8: When the indication parameter is abnormal, a new simulated product vector is input into the preset analysis model based on time series data, and the parameters of the simulated product vector are continuously adjusted until the obtained indication parameter is displayed as normal, thereby obtaining a target simulated product vector;

[0070] S9: extracting the product vector with the latest time sequence within the set time period as a reference product vector;

[0071] S10: Compare the reference product vector with the target simulated product vector, and set the marketing intensity of each product according to the comparison result.

[0072] As described in step S1 above, the review information is obtained from the preset review database through the preset crawler network. The review information is obtained from the preset review database through the preset crawler network. These review information may come from social media, e-commerce merchants, online forums, etc., and include consumers' direct feedback and potential demand for products. The use of crawler technology enables merchants to collect this data on a large scale and efficiently, providing rich materials for subsequent analysis.

[0073] As described in step S2 above, target sales information for various products is set based on the review information. Based on the collected review information, merchants can set target sales information for various products. These goals may include key indicators such as sales, market share, and customer satisfaction. By using AI technology to perform sentiment analysis, topic extraction, and other processing on these reviews, merchants can more accurately understand consumer needs and set goals that are more in line with market realities. In some embodiments, target sales information for various products can also be set manually based on statistics of review information.

[0074] As described in the above steps S3-S4, the sales information of various products at multiple time points within the time period set by the merchant is obtained, and the sales information of various products at each time point is vectorized to obtain the product vector corresponding to each time point. The sales information of various products at multiple time points within the time period set by the merchant is obtained. This information usually includes sales volume, sales revenue, return rate, etc. In order to input this sales information into the AI ​​model for analysis, it is necessary to vectorize it, that is, convert the sales information into a mathematical vector form for subsequent calculation and comparison.

[0075] As described in step S5 above, each of the product vectors is sequentially input into the preset analysis model based on time series data in chronological order to obtain the sales status vector of the merchant in the set time period, and the product vectors at each time point are sequentially input into the preset analysis model based on time series data in chronological order. This model can capture the trend and pattern of sales data changing over time, thereby predicting future sales status. In this way, merchants can more accurately understand the changing trends of market dynamics and consumer behavior. Specifically, the preset analysis model based on time series data is trained by using a large number of product vectors as the input of the model and the corresponding sales status vectors as the output of the model.

[0076] As described in the above steps S6-S7, the sales status vector is normalized to obtain corresponding indication parameters, the sales status vector is normalized to obtain corresponding indication parameters, and the indication parameters are compared with the set thresholds. These indication parameters can intuitively reflect the quality of the sales status. Next, the indication parameters are compared with the set thresholds to detect whether there are abnormalities. The set thresholds are usually obtained based on the sales information of various products at the first time point within the time period set by the merchant and the target sales information of various products. If the indication parameter exceeds the set threshold, it is determined to be abnormal, and it is necessary to further analyze the cause and take corresponding measures. The set threshold is a pre-set value.

[0077] As described in step S8 above, when the indication parameter is abnormal, a new simulated product vector is input into the preset analysis model based on time series data, and the parameters of the simulated product vector are continuously adjusted until the obtained indication parameter is displayed as normal, thereby obtaining the target simulated product vector. When the indication parameter is abnormal, a new simulated product vector is input into the analysis model based on time series data, and the parameters of the simulated product vector are continuously adjusted until the obtained indication parameter is displayed as normal. This process is actually an optimization process, which simulates different marketing strategies and product combinations to find the best solution to restore the sales status to normal. That is, the user needs to meet certain requirements at the next time point, so that the merchant can make adjustments based on this.

[0078] As described in the above steps S9-S10, extract the product vector with the latest timing in the set time period as the reference product vector, compare the reference product vector with the target simulation product vector, and set the marketing intensity of each product type according to the comparison result. Extract the product vector with the latest timing in the set time period as the reference product vector, and compare it with the target simulation product vector. According to the comparison results, merchants can set the marketing intensity of each product type. For example, for products with poor performance, you can increase marketing investment and publicity efforts; for products with good performance, you can maintain or adjust the marketing strategy appropriately. Merchants can adjust their marketing strategies based on review information, and thus maximize the promotion of their products, providing merchants with an intelligent marketing plan that can not only help merchants better understand market dynamics and consumer needs, but also find the best marketing strategy through simulation and optimization, thereby enhancing the competitiveness and market share of merchants.

[0079] In one embodiment, the preset analysis model based on time series data is a long short-term memory network model, and the step S5 of inputting each of the product vectors into the preset analysis model based on time series data in chronological order to obtain the sales state vector of the merchant within the set time period includes:

[0080] S501: Input each of the product vectors into the forget gate, input gate, new information, and output gate of each recurrent unit in the long short-term memory network in sequence;

[0081] S502: According to formula F t =sigmoid(W xf |X t |+W hf H t-1 +b f )

[0082] I t =sigmoid(W xi |X t |+W hi H t-1 +b i )

[0083] O t =sigmoid(W xo |X t |+W ho H t-1 +b o )

[0084] N t =tanh(W xn |X t |+W hn Ht-1 +b n )

[0085] C t =F t ⊙C t-1 +I t ⊙N t

[0086] S503: Calculate the output state of each cycle unit in turn, so as to obtain the sales state vector output by the last cycle unit; where F t represents the tth forget gate, I t represents the tth input gate, N t represents the tth new information, O t represents the tth output gate, C t represents the tth output state, X t represents the t-th product vector, W xf , W hf 、b f , W xi , W hi 、b i , W xo , W ho 、b o , W xn , W hn 、b n These are the preset parameters.

[0087] As described in the above steps S501-S503, the calculation of the sales state vector is realized. Specifically, in the long short-term memory network, since the state calculation of each step combines the state of the previous time step and the newly added information, the weight is determined by the input gate and the forget gate, that is, the bias of the merchant's previous sales products is obtained, so that the final sales state vector can be obtained.

[0088] In one embodiment, when the indication parameter is abnormal, a new simulated product vector is input into the preset analysis model based on time series data, and the parameters of the simulated product vector are continuously adjusted until the obtained indication parameter is displayed as normal, and step S8 of obtaining the target simulated product vector includes:

[0089] S801: When the indication parameter is abnormal, one or more recurrent units are added to the long short-term memory network, and a simulated product vector is input to the newly added recurrent unit and the simulated product vector is continuously adjusted until the indication parameter is normal, thereby obtaining a final adjusted target simulated product vector.

[0090] As described in the above step S801, the acquisition of the target simulated product vector is realized, that is, it is input into the simulated product vector and the simulated product vector is continuously adjusted until the indication parameters are normal. That is, after the merchant's marketing, the sales of its products should meet certain requirements to make the indication parameters normal, and according to the simulated product vector, a marketing direction can be provided for the merchant, so that the merchant can conduct marketing in this direction, providing guidance for the merchant.

[0091] In one embodiment, the step S1 of acquiring comment information from a preset comment database through a preset crawler network includes:

[0092] S101: Obtaining location information of the preset review database;

[0093] S102: Obtain corresponding comment information at the location corresponding to the location information through a sqoop script.

[0094] As described in the above steps S101-S102, since the comment database exists on different platforms, such as various social network platforms and sales platforms, the comment database can be obtained, and then the corresponding comment information on each platform can be obtained through the sqoop script.

[0095] In one embodiment, the step S2 of setting target sales information of various products according to the review information includes:

[0096] S201: Processing each of the review information according to a preset natural language processing model and sentiment analysis model to analyze the customer's purchase expectations;

[0097] S202: Setting target sales information for each product type according to the historical sales information and the purchase expectation value.

[0098] As described in the above steps S201-S202, the target sales information of various products is set. That is, the collected comment information is cleaned to remove irrelevant information, noise and repeated content. Basic NLP processing such as word segmentation and part-of-speech tagging is performed. Sentiment analysis: Use pre-trained sentiment analysis models (such as BERT, Transformer and other models based on deep learning) to analyze the sentiment tendency of each comment to determine whether it is positive, negative or neutral. Identify the sentiment intensity in the comments, that is, the degree of customer satisfaction or dissatisfaction with the product or service. Apply topic models (such as LDA, LSI, etc.) to identify the main topics or concerns mentioned in the comments, involving the product's functions, performance, price, appearance, after-sales service and other aspects. Combine the results of sentiment analysis and topic extraction to analyze customer expectations and satisfaction with various aspects of the product. Identify the product features that customers are most concerned about and the improvement points they expect. Review sales data over a period of time, including sales volume, sales revenue, return and exchange rates, etc. Analyze which product features or promotional strategies have a positive impact on sales and which have a negative impact. Combine the customer purchase expectations obtained from the analysis with historical sales information. Identify product features or combinations that meet customer expectations and have historically performed well in sales. Based on this analysis, set specific target sales information for each product category.

[0099] In one embodiment, after the step S7 of comparing the indication parameter with the set threshold, the method further includes:

[0100] S811: When the indicated parameters are normal, the previous marketing plan is continued to be executed.

[0101] As described in the above step S811, if the indication parameter is normal, it means that the marketing situation of the merchant is consistent with each review information, therefore, the previous marketing plan is successful and can continue to be executed according to the original plan.

[0102] In one embodiment, the step S10 of comparing the reference product vector with the target simulated product vector and setting the marketing strength of each product according to the comparison result includes:

[0103] S1001: Calculate a difference vector between the reference product vector and the target simulated product vector;

[0104] S1002: extracting first values ​​and second values ​​of various types of products from the difference vector and the reference product vector;

[0105] S1003: Compare the first value with the second value to obtain a ratio of each type of product;

[0106] S1004: Setting the marketing intensity of each product according to the ratio.

[0107] As described in the above steps S1001-S1004, the difference vector and the reference product vector are traversed to extract the first value and the second value of each product feature accordingly. For each product feature, the first value (difference) is divided by the second value (benchmark value) to obtain the ratio of the feature. This ratio reflects the degree of improvement or difference of the target simulation product in this feature relative to the market benchmark. Evaluate the ratio of each feature to determine which features have significant improvements and which features have smaller differences. For features with larger ratios, that is, the target simulation product has significant improvements or differences in these features, marketing efforts should be increased to emphasize these advantages. For features with smaller ratios, that is, the target simulation product is not much different from the market benchmark in these features, you can maintain conventional marketing efforts or reduce investment appropriately.

[0108] Reference Figure 2 The present invention provides a marketing full-link analysis device based on artificial intelligence, comprising:

[0109] A first acquisition module 10 is used to acquire comment information from a preset comment database through a preset crawler network;

[0110] A setting module 20, used to set target sales information of various products according to the review information;

[0111] The second acquisition module 30 is used to obtain sales information of various products at multiple time points within a time period set by the merchant;

[0112] A vectorization module 40 is used to vectorize the sales information of various products at various time points to obtain product vectors corresponding to various time points;

[0113] A first input module 50 is used to input each of the product vectors into a preset analysis model based on time series data in chronological order to obtain a sales status vector of the merchant within the set time period;

[0114] A normalization module 60, used to normalize the sales state vector to obtain a corresponding indication parameter;

[0115] A first comparison module 70 is used to compare the indication parameter with a set threshold value; wherein, when the linear distance is higher than the set threshold value, the indication parameter is determined to be abnormal, and the set threshold value is obtained based on the sales information of various products at the first time point within the time period set by the merchant and the target sales information of various products;

[0116] The second input module 80 is used to input a new simulated product vector into the preset analysis model based on time series data when the indication parameter is abnormal, and continuously adjust the parameters of the simulated product vector until the obtained indication parameter is displayed as normal, thereby obtaining a target simulated product vector;

[0117] An extraction module 90 is used to extract the product vector with the latest time sequence within the set time period as a reference product vector;

[0118] The second comparison module 100 is used to compare the reference product vector with the target simulated product vector, and set the marketing strength of each product according to the comparison result.

[0119] In one embodiment, the preset analysis model based on time series data is a long short-term memory network model, and the first input module 50 includes:

[0120] An input submodule, used to sequentially input each of the product vectors into a forget gate, an input gate, new information, and an output gate in each recurrent unit in the long short-term memory network;

[0121] The calculation submodule is used to calculate the t =sigmoid(W xf |X t |+W hf H t-1 +b f )

[0122] I t =sigmoid(W xi |X t |+W hi H t-1 +b i )

[0123] O t =sigmoid(W xo |X t |+W ho H t-1 +b o )

[0124] N t =tanh(W xn |X t |+W hn H t-1 +b n )

[0125] C t =F t ⊙C t-1 +I t ⊙Nt

[0126] The output state of each cycle unit is calculated in turn to obtain the sales state vector output by the last cycle unit; where F t represents the tth forget gate, I t represents the tth input gate, N t represents the tth new information, O t represents the tth output gate, C t represents the tth output state, X t represents the t-th product vector, W xf , W hf , b f , W xi , W hi , b i , W xo , W ho , b o , W xn , W hn , b n These are the preset parameters.

[0127] In one embodiment, the second input module 80 comprises:

[0128] A submodule is added to add one or more recurrent units to the long short-term memory network when the indication parameter is abnormal, and a simulated product vector is input into the newly added recurrent unit and the simulated product vector is continuously adjusted until the indication parameter is normal, thereby obtaining the final adjusted target simulated product vector.

[0129] In one embodiment, the first acquisition module 10 includes:

[0130] A location information acquisition submodule is used to acquire the location information of the preset comment database;

[0131] The comment information acquisition submodule is used to acquire corresponding comment information at the position corresponding to the position information through a sqoop script.

[0132] In one embodiment, the setting module 20 includes:

[0133] An analysis submodule, for processing each of the review information according to a preset natural language processing model and sentiment analysis model, and analyzing the customer's purchase expectations;

[0134] The setting submodule is used to set target sales information for various products based on historical sales information and the purchase expectation value.

[0135] In one embodiment, the marketing full-link analysis device based on artificial intelligence further includes:

[0136] The continuing execution module is used to continue execution according to the previous marketing plan when the indication parameters are normal.

[0137] In one embodiment, the second comparison module 100 includes:

[0138] A difference vector calculation submodule, used to calculate the difference vector between the reference product vector and the target simulated product vector;

[0139] A numerical value extraction submodule, used to extract the first numerical value and the second numerical value of each type of product in the difference vector and the reference product vector;

[0140] A comparison submodule, used to compare the first value with the second value to obtain a ratio of each type of product;

[0141] The setting submodule is used to set the marketing intensity of each type of product according to the ratio.

[0142] The beneficial effects of the present invention are as follows: With the help of an analysis model based on time series data, a large amount of sales data can be automatically processed and analyzed, reducing manual intervention and improving analysis efficiency. By normalization processing and threshold comparison, when indicating abnormal parameters, by inputting new simulated product vectors into the model and adjusting parameters, by comparing the reference product vector with the target simulated product vector, merchants can formulate personalized marketing strategies for each type of product to improve the pertinence and effectiveness of marketing. By accurately analyzing consumer demand and market trends, and formulating personalized marketing strategies, merchants can better meet consumer demand and enhance brand image and market competitiveness.

[0143] Reference Figure 3 In an embodiment of the present application, a computer device is also provided. The computer device may be a server, and its internal structure may be as follows: Figure 3 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. Among them, the processor designed by the computer is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store various sales information, etc. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the marketing full-link analysis method based on artificial intelligence described in any of the above embodiments can be implemented.

[0144] Those skilled in the art will understand that Figure 3The structure shown in is merely a block diagram of a portion of the structure related to the present application solution and does not constitute a limitation on the computer device to which the present application solution is applied.

[0145] An embodiment of the present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the artificial intelligence-based full-link marketing analysis method described in any of the above embodiments can be implemented.

[0146] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided in this application and used in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0147] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method 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, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, device, article or method including the element.

[0148] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0149] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0150] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention shall be included in the scope of the claims of the present invention.

Claims

1. A marketing full-link analysis method based on artificial intelligence, characterized in that: include: Obtaining comment information from a preset comment database through a preset crawler network; Setting target sales information for various products based on the review information; Obtain sales information of various products at multiple time points within the time period set by the merchant; Vectorize the sales information of various products at each time point to obtain the product vector corresponding to each time point; Inputting each of the product vectors into a preset analysis model based on time series data in chronological order, and obtaining a sales status vector of the merchant within the set time period; Normalizing the sales state vector to obtain corresponding indication parameters; Compare the indication parameter with a set threshold; wherein, when the linear distance is higher than the set threshold, the indication parameter is determined to be abnormal, and the set threshold is obtained based on the sales information of various products at the first time point within the time period set by the merchant and the target sales information of various products; When the indication parameter is abnormal, a new simulated product vector is input into the preset analysis model based on time series data, and the parameters of the simulated product vector are continuously adjusted until the obtained indication parameter is displayed as normal, thereby obtaining a target simulated product vector; Extracting the product vector with the latest timing within the set time period as a reference product vector; The reference product vector is compared with the target simulated product vector, and the marketing intensity of each product is set according to the comparison result.

2. The marketing full-link analysis method based on artificial intelligence according to claim 1, characterized in that: The preset analysis model based on time series data is a long short-term memory network model. The step of inputting each of the product vectors into the preset analysis model based on time series data in chronological order to obtain the sales state vector of the merchant within the set time period includes: Input each of the product vectors into the forget gate, input gate, new information, and output gate of each recurrent unit in the long short-term memory network in sequence; According to formula F t =sigmoid(W xf |X t |+W hf H t-1 +b f ) I t =sigmoid(W xi |X t |+W hi H t-1 +b i ) O t =sigmoid(W xo |X t |+W ho H t-1 +b o ) N t =tanh(W xn |X t |+W hn H t-1 +b n ) C t =F t ⊙C t-1 +I t ⊙N t The output state of each cycle unit is calculated in turn to obtain the sales state vector output by the last cycle unit; where F t represents the tth forget gate, I t represents the tth input gate, N t represents the tth new information, O t represents the tth output gate, C t represents the tth output state, X t represents the t-th product vector, W xf , W hf , b f , W xi , W hi , b i , W xo , W ho , b o , W xn , W hn , b n These are the preset parameters.

3. The marketing full-link analysis method based on artificial intelligence according to claim 2, characterized in that: When the indication parameter is abnormal, a new simulated product vector is input into the preset analysis model based on time series data, and the parameters of the simulated product vector are continuously adjusted until the obtained indication parameter is displayed as normal, and the step of obtaining the target simulated product vector includes: When the indication parameter is abnormal, one or more recurrent units are added to the long short-term memory network, and the simulated product vector is input into the newly added recurrent unit and the simulated product vector is continuously adjusted until the indication parameter is normal, thereby obtaining the final adjusted target simulated product vector.

4. The marketing full-link analysis method based on artificial intelligence according to claim 1, characterized in that: The step of obtaining comment information from a preset comment database through a preset crawler network includes: Obtaining location information of the preset comment database; The corresponding comment information is obtained at the position corresponding to the position information through the sqoop script.

5. The marketing full-link analysis method based on artificial intelligence according to claim 1, characterized in that: The step of setting target sales information of various products according to the review information includes: Process each of the review information according to a preset natural language processing model and sentiment analysis model to analyze the customer's purchase expectations; Target sales information for each product category is set based on historical sales information and the purchase expectations.

6. The marketing full-link analysis method based on artificial intelligence according to claim 1, characterized in that: After the step of comparing the indication parameter with the set threshold, the method further includes: When the indicated parameters are normal, the previous marketing plan is continued.

7. The marketing full-link analysis method based on artificial intelligence according to claim 1, characterized in that: The step of comparing the reference product vector with the target simulated product vector and setting the marketing strength of each product according to the comparison result includes: Calculating a difference vector between the reference product vector and the target simulated product vector; Extracting first and second values ​​of various types of products from the difference vector and the reference product vector; Compare the first value with the second value to obtain the ratio of each type of product; The marketing intensity of each product category is set according to the ratio.

8. A marketing full-link analysis device based on artificial intelligence, characterized in that: include: A first acquisition module is used to acquire comment information from a preset comment database through a preset crawler network; A setting module, used to set target sales information of various products according to the review information; The second acquisition module is used to obtain sales information of various products at multiple time points within a time period set by the merchant; The vectorization module is used to vectorize the sales information of various products at various time points to obtain the product vectors corresponding to each time point; A first input module, used to input each of the product vectors into a preset analysis model based on time series data in chronological order, to obtain a sales status vector of the merchant within the set time period; A normalization module, used for normalizing the sales state vector to obtain corresponding indication parameters; A first comparison module is used to compare the indication parameter with a set threshold value; wherein, when the linear distance is higher than the set threshold value, the indication parameter is determined to be abnormal, and the set threshold value is obtained based on the sales information of various products at the first time point within the time period set by the merchant and the target sales information of various products; A second input module is used for inputting a new simulated product vector into the preset analysis model based on time series data when the indication parameter is abnormal, and continuously adjusting the parameters of the simulated product vector until the obtained indication parameter is displayed as normal, thereby obtaining a target simulated product vector; An extraction module, used for extracting the product vector with the latest time sequence within the set time period as a reference product vector; The second comparison module is used to compare the reference product vector with the target simulated product vector, and set the marketing intensity of each product according to the comparison result.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images