Search recommendation word generation method and device, computer device, and storage medium

Abstract

The application relates to a search recommendation word generation method and device, computer equipment and a storage medium. The method involves cloud technology, including obtaining a historical search word carrying seasonal attribute information, obtaining a historical search period of the historical search word, and then determining a recommendation index corresponding to the historical search word according to the whole-network historical search heat data of the historical search word in the historical search period and the search times of the historical search word on a current application platform. According to the recommendation index and the seasonal attribute information, the generated search recommendation word list is adjusted and displayed. The method can flexibly adjust the search word recommendation list generated according to the recommendation index and the seasonal attribute information of the historical search word with the seasonal attribute information, so that the product search and purchase of a user can be positively recommended and guided at a suitable time point, invalid recommendation can be avoided, the correlation between the recommendation word and the user is improved, and the guiding effect and the recommendation effect brought by the recommendation word are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computers, and in particular to a search recommendation word generation method and device, computer equipment and a storage medium. BACKGROUND

[0002] With the development of computer technology and the widespread use of Internet technology in people's work and life, more and more users use the Internet for information acquisition and transmission, and purchase products through e-commerce platforms. When purchasing products on e-commerce platforms, users usually search for the required products through the search entry, and further select the actual required products according to the search results. Among them, the search entry of the e-commerce platform, as a channel for users to quickly reach the required goods, plays an important role in guiding users, and the search results corresponding to the search entry usually contain association words or suggestion words, which can effectively help users complete the search process faster and find the required products to achieve transactions.

[0003] Traditionally, the association word is obtained based on the portrait information of the use object or the product title input by the user, for example, if a user searches for "air conditioner" recently, the association word "air conditioner" is pushed to the user subsequently, or the association word is obtained by the number of keywords in the product library or the keywords of the newly imported products in recent period, and the association word is displayed around the search input box, and if the user clicks, the user can directly jump to the search result page of the corresponding search word to view the specific search results.

[0004] However, the traditional association word generation method, such as the recommendation word mining based on the behavior data of the use object, is prone to interest delay and heat dissipation, so that users who search and purchase "XX product" do not get effective search guidance in the early stage, and when these users improve the recommendation coefficient of "XX product" through search as a behavior log, the purchase peak period has passed, and a large amount of purchase demand has fallen, at this time, as an association word, it cannot have a good guiding effect and recommendation effect. Therefore, the association word generated by the traditional method still has the problem of insufficient association with the user and poor recommendation effect. SUMMARY

[0005] Therefore, it is necessary to provide a search recommendation word generation method, device, computer equipment and storage medium capable of improving the association between the recommendation words displayed by the search entry and the user, and the guiding effect and recommendation effect brought by the recommendation words displayed by the search entry.

[0006] A search recommendation word generation method, the method comprising:

[0007] acquire historical search words carrying seasonal attribute information, and acquire a historical search period of the historical search words;

[0008] determine a recommendation index corresponding to the historical search words according to historical search heat data of the historical search words in the historical search period and a search frequency of the historical search words on a current application platform;

[0009] adjust and generate a search recommendation word list according to the recommendation index and the seasonal attribute information, and display the search recommendation word list.

[0010] A search recommendation word generation device, the device comprising:

[0011] a historical search word acquisition module configured to acquire historical search words carrying seasonal attribute information, and acquire a historical search period of the historical search words;

[0012] a recommendation index determination module configured to determine a recommendation index corresponding to the historical search words according to historical search heat data of the historical search words in the historical search period and a search frequency of the historical search words on a current application platform;

[0013] a search recommendation word list generation module configured to adjust and generate a search recommendation word list according to the recommendation index and the seasonal attribute information, and display the search recommendation word list.

[0014] A computer device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the following steps when executing the computer program:

[0015] acquire historical search words carrying seasonal attribute information, and acquire a historical search period of the historical search words;

[0016] determine a recommendation index corresponding to the historical search words according to historical search heat data of the historical search words in the historical search period and a search frequency of the historical search words on a current application platform;

[0017] adjust and generate a search recommendation word list according to the recommendation index and the seasonal attribute information, and display the search recommendation word list.

[0018] A computer readable storage medium having a computer program stored thereon, the computer program being executed by a processor to implement the following steps:

[0019] acquire historical search words carrying seasonal attribute information, and acquire a historical search period of the historical search words;

[0020] determine a recommendation index corresponding to the historical search word according to the historical search word's whole-network historical search heat data in the historical search period and the search times of the historical search word on the current application platform;

[0021] generate and display the search recommendation word list according to the recommendation index and the seasonal attribute information.

[0022] A computer program product comprising a computer program which, when executed by a processor, implements the following steps:

[0023] obtain a historical search word carrying seasonal attribute information and a historical search period of the historical search word;

[0024] determine a recommendation index corresponding to the historical search word according to the historical search word's whole-network historical search heat data in the historical search period and the search times of the historical search word on the current application platform;

[0025] generate and display the search recommendation word list according to the recommendation index and the seasonal attribute information.

[0026] In the search recommendation word generation method, device, computer equipment and storage medium, by obtaining a historical search word carrying seasonal attribute information and a historical search period of the historical search word, a recommendation index corresponding to the historical search word is determined according to the historical search word's whole-network historical search heat data in the historical search period and the search times of the historical search word on the current application platform. Then, the search recommendation word list is generated and displayed according to the recommendation index and the seasonal attribute information. The historical search word with seasonal attribute information can be flexibly adjusted according to the recommendation index and the seasonal attribute information to realize positive recommendation and guide users to search and purchase products at the right time, avoid invalid recommendation, and improve the relevance between the recommendation words displayed on the application platform and the users and the guiding effect and recommendation effect of the recommendation words. BRIEF DESCRIPTION OF DRAWINGS

[0027] Figure 1 An application environment diagram of the search recommendation word generation method in one embodiment;

[0028] Figure 2 A flowchart of the search recommendation word generation method in one embodiment;

[0029] Figure 3 A search trend diagram of "moon cakes" in one embodiment;

[0030] Figure 4Fig. 1 is a diagram showing the search trend of "crayfish" in one embodiment;

[0031] Figure 5 Fig. 2 is a diagram showing the comparison of the search recommendation word list before and after adjustment in one embodiment;

[0032] Figure 6 Fig. 3 is a diagram showing the process of determining whether each historical search word carries seasonal attribute information in one embodiment;

[0033] Figure 7 Fig. 4 is a diagram showing the moving average heat trend of "crayfish" in one embodiment;

[0034] Figure 8 Fig. 5 is a diagram showing the moving average heat trend of "snacks" in one embodiment;

[0035] Figure 9 Fig. 6 is a diagram showing the current average heat index trend of "crayfish" in one embodiment;

[0036] Figure 10 Fig. 7 is a diagram showing the current average heat index trend of "snacks" in one embodiment;

[0037] Figure 11 Fig. 8 is a diagram showing the process of determining whether each historical search word carries seasonal attribute information in another embodiment;

[0038] Figure 12 Fig. 9 is a diagram showing the process of generating search recommendation word in yet another embodiment;

[0039] Figure 13 Fig. 10 is a diagram showing the historical search heat trend of real-time hot key word in one embodiment;

[0040] Figure 14 Fig. 11 is a diagram showing the process of generating search recommendation word in still another embodiment;

[0041] Figure 15 Fig. 12 is a block diagram showing the structure of the search recommendation word generating device in one embodiment;

[0042] Figure 16 Fig. 13 is a diagram showing the internal structure of the computer device in one embodiment. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0044] The search recommendation word generation method provided in the application relates to cloud technology. The cloud technology refers to a kind of hosting technology that unifies a series of resources such as hardware, software, network in a wide area network or local area network, realizes the calculation, storage, processing and sharing of data. The cloud technology is based on the network technology, information technology, integration technology, management platform technology, application technology and the like applied in the cloud computing business model, can form a resource pool, and is used on demand, flexible and convenient. Cloud computing technology will become an important support. The background service of technical network system needs a large amount of calculation and storage resources, such as video websites, picture websites and more portal websites. With the high development and application of the Internet industry, in the future, every item may have its own identification mark, and needs to be transmitted to the background system for logical processing. Different levels of data will be processed separately, and various industry data need strong system backup support, which can only be realized through cloud computing.

[0045] Cloud computing is a computing mode that distributes computing tasks on a resource pool composed of a large number of computers, so that various application systems can obtain computing power, storage space and information services according to needs. The network providing resources is called "cloud". The resources in the "cloud" can be infinitely expanded in the eyes of the user, and can be obtained at any time, used on demand, expanded at any time, and paid according to use.

[0046] As a basic capability provider of cloud computing, a cloud computing resource pool (referred to as a cloud platform, generally referred to as an IaaS (Infrastructure as a Service) platform) is established, a plurality of types of virtual resources are deployed in the resource pool, and external customers can select and use them. The cloud computing resource pool mainly includes: computing devices (virtualized machines containing operating systems), storage devices, network devices. According to logical functions, the PaaS (Platform as a Service) layer can be deployed on the IaaS (Infrastructure as a Service) layer, and the SaaS (Software as a Service) layer is deployed on the PaaS layer. SaaS can also be directly deployed on IaaS. PaaS is a platform for software running, such as databases, web containers, etc. SaaS is various business software, such as web portal websites, SMS mass senders, etc. Generally, SaaS and PaaS are upper layers relative to IaaS.

[0047] The search recommendation word generation method provided in the application specifically relates to cloud computing in cloud technology, and can be applied to, for example Figure 1The application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data required by the server 104 to process, and the data storage system can be integrated on the server 104, or placed on the cloud or other network servers. Among them, the server 102 obtains the historical search words carrying the seasonal attribute information, and obtains the historical search period of the historical search words. Among them, the historical search words carrying the seasonal attribute information and the historical search period of the historical search words can be obtained from the local storage of the terminal 102, or from the data storage system of the server 104 itself, or from the data storage system of other network servers. Further, the server 104 can determine the recommendation index corresponding to the historical search words according to the historical search heat data of the historical search words in the historical search period and the search times of the historical search words on the current application platform. Further, the server 104 adjusts the search recommendation word list generated according to the recommendation index and the seasonal attribute information, and displays the search recommendation word list on the application program of the terminal or the search website. Among them, the terminal 102 can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, a vehicle terminal, a smart television, etc., but is not limited thereto. The server 104 can be a standalone physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDNs, and big data and artificial intelligence platforms. The terminal 102 and the server 104 can be directly or indirectly connected through wired or wireless communication, which is not limited in the present application.

[0048] In one embodiment, as shown in Figure 2 , a search recommendation word generation method is provided, which is applied to the server in Figure 1 for example, including the following steps:

[0049] Step S202, obtaining historical search words carrying seasonal attribute information, and obtaining the historical search period of the historical search words.

[0050] Specifically, the historical search words carrying the seasonal attribute information are obtained from the current application platform and the whole network search platform, and the historical search period of each historical search word is obtained. Among them, the seasonal attribute information represents the specific information of the search word carrying the season, festival and solar term, which can be a specific holiday or season, such as summer, winter, Mid-Autumn Festival, Dragon Boat Festival, summer solstice or winter solstice, etc. It can also be the corresponding specific time point of different seasons, solar terms and festivals, such as Mid-Autumn Festival on the 15th day of the eighth lunar month, Dragon Boat Festival on the 5th day of the fifth lunar month, etc.

[0051] Corresponding to each of the historical search terms carrying seasonal attribute information, there is a corresponding search period. For example, the search period of the search terms such as “Mid-Autumn” and “Moon Cake” corresponds to the period before the Mid-Autumn Festival each year. The search period of the historical search terms can be determined according to the search frequency and the search frequency of the historical search terms carrying seasonal attribute information obtained from the current application platform and the search platform of the entire network. For example, the search frequency and the search frequency of the search terms such as “Mid-Autumn” and “Moon Cake” are greater than those in other periods of the same year from the first day of August to the fifteenth day of August each year. Therefore, it can be determined that the historical search period of the search terms such as “Mid-Autumn” and “Moon Cake” is from the first day of August to the fifteenth day of August each year.

[0052] In one embodiment, before obtaining the historical search terms carrying seasonal attribute information and obtaining the historical search period of the historical search terms, the method further comprises:

[0053] Collecting historical search data of the current application platform and the search engine of the entire network in a first preset time period; traversing each historical search data to extract historical search terms associated with a time period; and performing seasonal attribute analysis on the historical search terms associated with the time period to determine whether each historical search term carries seasonal attribute information.

[0054] Specifically, by collecting historical search data of the current application platform and the search engine of the entire network in a first preset time period, for example, when the current application platform is an e-commerce shopping platform, historical search data of all users of the current application platform in the past ten years or the past five years can be obtained, and historical search data of all users of the search engine of the entire network in the corresponding time length can be obtained from a third-party platform.

[0055] Further, by traversing each historical search data, historical search terms associated with a time period are extracted from the historical search data, for example, historical search terms corresponding to specific time or time range of seasons, festivals and solar terms are extracted, for example, historical search terms corresponding to specific time of specific festivals or solar terms (Mid-Autumn Festival, Dragon Boat Festival and Winter Solstice, etc.). Further, seasonal attribute analysis is performed on the historical search terms associated with the time period to determine whether each historical search term carries seasonal attribute information.

[0056] In one embodiment, as shown in Figure 3 and Figure 4 respectively, the search trends of “Moon Cake” in previous years and the search trends of “Small Lobster” in previous years are provided. As shown in Figure 3 , for “Moon Cake”, the search request volume rapidly increases before the Mid-Autumn Festival each year, and rapidly decreases in one or two days after the festival, and the decreasing speed is much faster than the increasing speed. Similarly, as shown in Figure 4It can be seen that for "crayfish", the search popularity gradually increases in May every year and remains high for more than one month, so the historical search period of "crayfish" is determined to be from May to June every year.

[0057] Referring to Figure 3 and Figure 4 It can be seen that "moon cake" and "crayfish" have corresponding historical search periods every year, and the search periods of different years are relatively small. By analyzing the seasonal attributes, it can be determined that "moon cake" and "crayfish" carry seasonal attribute information.

[0058] In step S204, the recommendation index corresponding to the historical search word is determined according to the historical search popularity data of the historical search word in the historical search period and the search times of the historical search word on the current application platform.

[0059] Specifically, the first adjustment parameter corresponding to the search times is determined according to the search times of the historical search word on the current application platform, and the recommendation index corresponding to the historical search word is further determined according to the historical search popularity data of the historical search word in the historical search period, the search times of the historical search word on the current application platform, and the first adjustment parameter.

[0060] Among them, by obtaining the historical search popularity data of the historical search word in the historical search period, for example, for the historical search word "crayfish", the historical search popularity data of the word in May of each year is obtained, including the historical search popularity data of the current application platform and the search engine in the whole network.

[0061] Similarly, the search times of the historical search word on the current application platform also need to be obtained, such as the search times of "crayfish" by all users on a certain e-commerce shopping platform, and the first adjustment parameter corresponding to the search times is further determined according to the search times of the historical search word on the current application platform.

[0062] Further, the value of the first adjustment parameter is used to represent the influence of the current search times on the overall recommendation weight. The more search times of the current application platform for the historical search word, the larger the first adjustment parameter. For the scene with less search sample, such as the search of CPS commission goods (the commission goods provided by each e-commerce platform to the outside, corresponding to a unique promotion link, if the link is clicked through external flow and transaction is achieved, a certain proportion of commission is made), since the order of goods is large but the search record is not much, the first adjustment parameter can be set to be small accordingly.

[0063] In one embodiment, the recommendation index is calculated by using the following formula (1):

[0064] ; (1)

[0065] P = W1 + a * count, wherein P is the recommendation index, W1 is the historical search hotness data of the historical search word in the historical search period, count is the search times of the historical search word on the current application platform, and a is the first adjustment parameter. Specifically, the product of the search times count of the historical search word on the current application platform and the first adjustment parameter a is calculated, and the product and the historical search hotness data W1 of the historical search word in the historical search period are summed up, and finally the recommendation index P corresponding to the historical search word is obtained.

[0066] Step S206, according to the recommendation index and the seasonal attribute information, the generated search recommendation word list is adjusted and displayed.

[0067] Specifically, according to the recommendation index, the search hotness value of each historical search word is determined, and the corresponding search word list is generated according to the search hotness value of each historical search word. Then, according to the seasonal attribute information, the corresponding search word list is adjusted in real time in the display time period corresponding to the historical search period, and is displayed.

[0068] Further, according to the recommendation index, the search hotness value of each historical search word can be determined, and the search hotness value of each historical search word is sorted according to the size. At the same time, the search word input by the user is also considered, and then the search recommendation word list generated according to the search word input by the user is adjusted according to the seasonal attribute information, that is, the historical search word carrying the seasonal attribute information is added to the search word list, and the display position of the historical search word in the search word list is adjusted according to the seasonal attribute information, and then the adjusted search recommendation word list is displayed around the search entrance.

[0069] Wherein, when adjusting the display position of the historical search word in the search word list according to the seasonal attribute information, the specific time and the search period of the historical search word need to be considered, for example, the current is April, and the search period of the historical search word "small lobster" is from May to July every year. Therefore, from April, the display position of "small lobster" in the search word list can be continuously improved, and after July, the display position of "small lobster" in the search word list can be reduced, so as to achieve real-time recommendation of seasonal goods and avoid invalid recommendation.

[0070] In an embodiment, taking "crayfish" as an example, according to the recommendation index and the seasonal attribute information (the search period of crayfish is from May to June every year), the recommendation weight of the associated word "crayfish" can be continuously increased from April 20 every year, the display position of "crayfish" in the search recommendation word list is gradually improved, and the recommendation weight of "crayfish" is decreased from July 1 of the same year, that is, the display position of "crayfish" in the search recommendation word list is adjusted downward, so as to dynamically adjust the search recommendation word list in real time, and achieve the recommendation effect of seasonal products.

[0071] In an embodiment, as shown in Figure 5 , a comparison diagram before and after adjustment of the search recommendation word list is provided, wherein, Figure 5 Fig. (a) in Figure 5 is used to represent a traditional recommendation list diagram before adjustment of the search word list, Fig. (b) in

[0072] is used to represent the search word list obtained after adjustment according to the recommendation index and the seasonal attribute information. Figure 5 Specifically, referring to Fig. (a) in , in the traditional search recommendation word list, a plurality of recommendation words are usually generated for user selection according to the search records of each user of the current application platform and the number of corresponding products, for example, if the user inputs "xiao", the traditional search recommendation word list can include "xiaomi, small refrigerator, small fan, small degree, small skin, small bear, xiaomi 11" and the like. Referring to Fig. (b) in Fig. (5), according to the recommendation index and the seasonal attribute information, the recommendation weight of "crayfish" in the search recommendation word list is improved, and the display position of "crayfish" in the search recommendation word list is improved. The adjusted search recommendation word list can include "xiaomi, crayfish, small fan, small air conditioner, small skin, small bear" and the like.

[0073] The above search recommendation word generation method can determine the recommendation index corresponding to the historical search word according to the whole network historical search heat data of the historical search word in the historical search period and the search times of the historical search word on the current application platform by acquiring the historical search word carrying the seasonal attribute information and acquiring the historical search period of the historical search word. Then, the generated search recommendation word list is adjusted and displayed according to the recommendation index and the seasonal attribute information. For the historical search word with seasonal attribute information, the generated search word recommendation list can be flexibly adjusted according to the recommendation index and the seasonal attribute information, so as to positively recommend and guide the user to search and purchase products at the appropriate time point, avoid invalid recommendation, and improve the association between the recommendation words displayed by the application platform and the user, and the guiding effect and recommendation effect brought by the recommendation words.

[0074] In an embodiment, as shown in Figure 6As shown, the step of determining whether each historical search term carries seasonal attribute information, i.e., the step of performing seasonal attribute analysis on the historical search terms associated with the time period, specifically includes:

[0075] In step S602, the occurrence time points of each historical search term are obtained, and search period analysis is performed based on the occurrence time points of each historical search term to generate corresponding historical search periods.

[0076] Specifically, by obtaining the occurrence time points of each historical search term and the number of occurrences at each occurrence time, and based on the occurrence time points and the number of occurrences of each historical search term, search period analysis is performed to determine the historical search period of each historical search term.

[0077] Further, taking "crayfish" as an example, by obtaining the search occurrence time points of the historical search term "crayfish" and the search frequency at each occurrence time point, such as "crayfish" appearing in April of each year and lasting until July of the same year, by obtaining the search frequency of "crayfish" every day from April to July of each year, the historical search period of "crayfish" is further determined. For example, the search frequency of "crayfish" from April to May is significantly less than the search frequency from May to June and from June to July of the same year, so the historical search period of "crayfish" can be further determined as from May to July of each year.

[0078] In step S604, the total network search frequency of each historical search term within the corresponding historical search period is extracted.

[0079] Specifically, by obtaining the total network search frequency of each historical search term within its own historical period on the current application platform (such as each e-commerce shopping platform) and the total network search engine.

[0080] In step S606, based on the total network search frequency, the current average heat index of the corresponding historical search term is determined.

[0081] Wherein, since the search keyword is extracted to increase the recommendation weight corresponding to the time, first of all, it needs to be clear whether the historical search term is strongly related to the season and time, i.e., it is further necessary to determine the current average heat index of the historical search term. For example, "moon cake" is related to the "mid-autumn festival" of the lunar calendar, and "crayfish" is related to the date of the solar calendar (such as from May to July of each year).

[0082] Specifically, the moving average of the total network search frequency is processed to generate the moving average heat data of the corresponding historical search term, and the annual average heat data of the historical search term is further obtained to generate the current average heat index of the historical search term according to the moving average heat data and the annual average heat data.

[0083] In one embodiment, based on the total network search times of each historical search word, the search times can be processed by moving average in time sequence, wherein the moving average processing means that a series of average numbers obtained by using the item-by-item progressive method to perform arithmetic average on several data items in the time sequence, if the number of average data items is N, it is called N period (item) moving average, the following formula (2) is used to perform moving average processing to obtain the moving average heat data SMAt of the historical search word:

[0084] (2)

[0085] Wherein, n represents the time period, Pn represents the search times of the nth day in the period, and SMAt represents the moving average heat data. In addition, weighted moving average processing or exponential moving average processing can be used to calculate the average of the search times to avoid the influence of individual extreme values on the normal reflection of the overall trend.

[0086] For example, the following table 1 is the calculation result of the moving average heat data when n=5:

[0087] Table 1

[0088]

[0089] In one embodiment, as shown in Figure 7 , the moving average heat trend of "crayfish" is provided, and as shown in Figure 7 , it can be seen that the search trend and the moving average heat trend of "crayfish" are specifically included, and based on the total network search times of "crayfish", i.e. the search trend, the moving average processing is performed to obtain the corresponding moving average heat trend. Based on the moving average heat trend of "crayfish", it can be determined that "crayfish" has an obvious time period, and the search heat is high in the corresponding time period.

[0090] In one embodiment, as shown in Figure 8 , the moving average heat trend of "snacks" is provided, and the search trend and the moving average heat trend of "snacks" are specifically included, and based on the total network search times of "snacks", i.e. the search trend, the moving average processing is performed to obtain the corresponding moving average heat trend. Based on the moving average heat trend of "snacks", it can be determined that "snacks" have purchase demand all year round, and do not have a specific search or purchase period, i.e. do not carry seasonal attribute information.

[0091] In one embodiment, the search times are moving averaged according to time sequence to obtain the moving average heat data of the historical search words, and the annual average heat data of the historical search words is further obtained. In order to further determine the corresponding high heat time, while avoiding the order of magnitude difference of heat, such as the whole network search times of "small crayfish" is between 0 and 20000, and the whole network search times of snack is between 0 and 2200, there is an order of magnitude difference between the two, which cannot be compared directly by the moving average heat trend of "small crayfish" and the moving average heat trend of "snack" shown in Figure 7 and Figure 8 It can be seen that the moving average heat trend of "small crayfish" and the moving average heat trend of "snack" are compared intuitively, and then the moving average heat data is further divided by the annual average heat data to obtain the current average heat index of the corresponding historical search word.

[0092] Further, as shown in Figure 9 and Figure 10 The current average heat index trend of "small crayfish" and the current average heat index trend of "snack" are provided respectively, and it can be known from Figure 9 that the current average heat index of "small crayfish" has obvious time periodicity, such as in May to July in 18 years, and in May to July in 19 years, and in May to July in 20 years, there is a heat rise, that is, "small crayfish" carries corresponding seasonal attribute information and has specific search period. It can be known from Figure 10 that "snack" has search demand and purchase demand all year round and does not have time periodicity.

[0093] Step S608, according to the current average heat index and the preset heat threshold, the seasonal attribute analysis of the historical search word is carried out to determine whether each historical search word carries seasonal attribute information.

[0094] Specifically, by obtaining the preset heat threshold, the current average heat index and the preset heat threshold are compared to determine whether there is a time point greater than the preset heat threshold in the current average heat index, and the time point greater than the preset heat threshold in the current average heat index is determined as the heat index vertex.

[0095] Further, the occurrence time of each heat index vertex, the time interval between each occurrence time, and the preset interval time threshold corresponding to the time interval between each occurrence time are obtained, and whether the time interval between the occurrence time of each heat index vertex in the past years is less than the preset interval time threshold is determined to analyze the seasonal attribute of the historical search word.

[0096] When it is determined that the time interval between the occurrence time of each heat index vertex in the past years is less than the preset interval time threshold, it is indicated that the corresponding historical search word carries seasonal attribute information.

[0097] In this embodiment, the time points of each historical search word are obtained, and the search cycle analysis is performed based on the time points of each historical search word to generate the corresponding historical search cycle. The number of searches of each historical search word in the corresponding historical search cycle is extracted, and the current average heat index of the corresponding historical search word is determined based on the number of searches. Then, the seasonal attribute analysis is performed on the historical search word according to the current average heat index and the preset heat threshold, to determine whether the historical search word carries seasonal attribute information. The seasonal attribute of the historical search word is determined from different angles to more accurately determine the search cycle of each historical search word and whether it carries seasonal attribute information, so as to adjust the recommendation weight of the historical search word carrying seasonal attribute information in the subsequent process, and provide more relevant guidance and recommendation when the user searches.

[0098] In one embodiment, as shown in Figure 11 The step of determining whether each historical search word carries seasonal attribute information, i.e., the step of performing seasonal attribute analysis on the historical search word according to the current average heat index and the preset heat threshold to determine whether each historical search word carries seasonal attribute information, specifically includes:

[0099] In step S1102, the heat index peak greater than the preset heat threshold is determined according to the current average heat index and the corresponding preset heat threshold.

[0100] Specifically, the preset heat threshold is obtained, and the preset heat threshold and the index of each time point in the current average heat index are compared to determine the time point of the index greater than the preset heat threshold as the heat index peak. The preset heat threshold can be adjusted and modified according to different application scenarios and actual conditions.

[0101] Further, taking "small lobster" as an example, the preset heat threshold can be 2.0. When the current average heat index is greater than the threshold preset heat threshold, it indicates that there is a case of rapid increase of the historical search word "small lobster" at a certain time. The time points corresponding to all average heat indexes greater than the threshold preset heat threshold are further selected to determine the heat index peak. Specifically, the heat index peaks of "small lobster" are [2018 / 5 / 5, 2019 / 5 / 10, 2020 / 5 / 13, 2021 / 5 / 22]. The search word "snack" does not satisfy the related points, and the overall fluctuation is small over time.

[0102] In step S1104, the time points of the heat index peaks in the past years are obtained, and the time intervals between each time point are determined.

[0103] Specifically, by acquiring the whole network historical search heat data of the historical search words in each year, and acquiring the occurrence time of the heat index vertex from the whole network historical search heat data, the time interval between the occurrence time of each heat index vertex in each year is determined.

[0104] Further, taking "crayfish" as an example, by acquiring the whole network historical search heat data of "crayfish" in each year, and acquiring the occurrence time of the heat index vertex from the whole network historical search heat data, for example, the occurrence time of the heat index vertex of "crayfish" in each year is mostly in May, such as May 5, May 10, May 13, etc., and further calculating the time interval between the occurrence time of the heat index vertex in each year, such as the time interval between May 5 and May 10, the time interval between May 10 and May 13, or the time interval between May 5 and May 13.

[0105] Step S1106, according to the time interval between the occurrence time and the corresponding preset interval time threshold, the seasonal attribute analysis of the historical search words is carried out, and whether each historical search word carries seasonal attribute information is judged.

[0106] Specifically, by comparing the time interval of the occurrence time of the heat index vertex with the preset interval time threshold, it is determined whether the time interval of the occurrence time of the heat index vertex is less than the preset interval time threshold.

[0107] Further, when the time interval of the occurrence time of the heat index vertex is less than the preset interval time threshold, it is considered that the corresponding historical search word has a periodic relationship. The preset interval time threshold can be adjusted and modified according to different application scenarios and actual needs.

[0108] For example, if the preset interval time threshold is 10, it means that the time points of the annual "crayfish" search peak appear within 10 days, and it is considered that there is a periodic increase in demand for "crayfish" in the middle of May each year. Since there are historical search words and lunar date relationship, such as the Mid-Autumn Festival, the Dragon Boat Festival, etc., when calculating the number of days relative to the first day of the year, the lunar calendar and the solar calendar need to be calculated at the same time. If the difference between the lunar calendar and the solar calendar is less than the preset interval time threshold, it is considered that the historical search word "crayfish" carries seasonal attribute information.

[0109] In this embodiment, according to the current average heat index and the corresponding preset heat threshold, the heat index vertex greater than the preset heat threshold is determined, and the occurrence time of the heat index vertex in the past years is further obtained to determine the time interval between each occurrence time. Then, according to the time interval between the occurrence times and the corresponding preset interval time threshold, the seasonal attribute analysis of the historical search words is performed to determine whether each historical search word carries seasonal attribute information. The seasonal attribute of the historical search words is determined from different angles to more accurately determine the search period of each historical search word and whether it carries seasonal attribute information, so as to subsequently adjust the recommendation weight of the historical search words carrying seasonal attribute information, and then provide more relevant guidance and recommendation when the user searches.

[0110] In one embodiment, as shown in Figure 12 , a search recommendation word generation method is provided. As shown in Figure 12 , the search recommendation word generation method specifically includes the following steps:

[0111] Step S1202, traversing each historical search data to obtain real-time hot keywords.

[0112] Specifically, by traversing each historical search data, the full-text hot event is analyzed, and the real-time hot keywords are extracted. For example, on July 23, 2021, “XX brand” has a large number of search and purchase behaviors due to real-time hotness. Therefore, by capturing the search heat of the entire network, the recommendation granularity of the corresponding suggestion words and association words is increased to guide the user's purchase demand for the products of the brand.

[0113] In one embodiment, as shown in Figure 13 , a historical search heat trend of real-time hot keywords is provided. As shown in Figure 13 , “XX brand” has high search heat and purchase heat from July 23 to July 28.

[0114] Step S1204, performing named entity recognition processing on each real-time hot keyword to extract product keywords matching product attributes.

[0115] Specifically, by performing named entity recognition processing (Named Entity Recognition processing) on each real-time hot keyword, the entity class, time class and number class in the text to be processed are recognized, such as: name, commodity, organization name, place name, time and other named entities, and then product keywords matching product attributes are extracted.

[0116] Step S1206, obtaining the historical search heat of each product keyword in a second preset time period and the search heat growth multiple.

[0117] Specifically, by acquiring product keywords such as "XX brand", "XX product" and the like, the historical search heat in a second preset time period and the search heat growth multiple in the second preset time period are obtained. The second preset time period can be adjusted and modified according to actual application scenarios and requirements. In this embodiment, the second preset time period is 30 days, and the historical search heat and the search heat growth multiple of the product keywords in the second preset time period, i.e. 30 days, can be obtained.

[0118] In step S1208, the second adjustment parameter is determined according to the search heat growth multiple.

[0119] Specifically, the second adjustment parameter associated with the search heat growth multiple is further determined according to the search heat growth multiple. The greater the search heat growth multiple, the greater the value of the second adjustment parameter, and the second adjustment parameter can also be adjusted and modified according to actual application scenarios and requirements.

[0120] In step S1210, the recommendation index of the product keyword is determined according to the historical search heat, the search heat growth multiple and the second adjustment parameter corresponding to the search heat growth multiple.

[0121] Specifically, according to the historical search heat, the search heat growth multiple and the second adjustment parameter corresponding to the search heat growth multiple, the recommendation index of the product keyword is calculated by using the following formula (3):

[0122] ; (3)

[0123] Wherein, Pr is the recommendation index of the product keyword, W2 is the historical search heat, X is the search heat growth multiple, and β is the second adjustment parameter. Specifically, the product of the search heat growth multiple X and the second adjustment parameter β of the product keyword in the second preset time period is calculated, and the sum of the product and the historical search heat W2 of the product keyword in the second preset time period is calculated, and finally the recommendation index Pr corresponding to the product keyword is obtained.

[0124] In step S1212, based on the recommendation index corresponding to each product keyword, a real-time recommendation keyword list is generated and displayed.

[0125] Specifically, after pulling the hot real-time point event, the recommendation index of the core product keyword of the hot real-time point event is improved, and then the order of recommendation is determined according to the recommendation index, and then the real-time recommendation keyword list is obtained and displayed around the search entrance of the application platform.

[0126] In the search recommendation word generation method, real-time hot keywords are obtained by traversing each historical search data, and named entity recognition processing is performed on each real-time hot keyword to extract product keywords matched with product attributes. Then, the historical search heat of each product keyword in a second preset time period and the search heat growth multiple are obtained, and the corresponding second adjustment parameter is determined according to the search heat growth multiple. The recommendation index of the product keyword is determined according to the historical search heat, the search heat growth multiple, and the second adjustment parameter corresponding to the search heat growth multiple, and a corresponding real-time recommendation word list is generated and displayed based on the recommendation index corresponding to each product keyword. Real-time hot events are used to extract product keywords, and the recommendation index of the product keyword is improved to provide more accurate recommendation guidance for users and further promote transaction completion.

[0127] In one embodiment, as shown in Figure 14 , a search recommendation word generation method is provided, which can be known from Figure 14 that the search recommendation word generation method specifically includes:

[0128] 1) Collect historical search data of a current application platform and a global search engine in a first preset time period.

[0129] 2) Traverse each historical search data to extract historical search words associated with a time period.

[0130] 3) Obtain the occurrence time point of each historical search word, and perform search period analysis based on the occurrence time point of each historical search word to generate a corresponding historical search period.

[0131] 4) Extract the global search number of each historical search word in the corresponding historical search period.

[0132] 5) Perform moving average processing on the global search number to generate moving average heat data of the corresponding historical search word.

[0133] 6) Obtain the annual average heat data of the historical search word, and generate the current average heat index of the historical search word according to the moving average heat data and the annual average heat data.

[0134] 7) Determine the heat index vertex greater than the preset heat threshold according to the current average heat index and the corresponding preset heat threshold.

[0135] 8) Obtain the occurrence time of the heat index vertex in the past years, and determine the time interval between each occurrence time.

[0136] 9) Perform seasonal attribute analysis on the historical search word according to the time interval between the occurrence times and the corresponding preset interval time threshold to determine whether the historical search word carries seasonal attribute information.

[0137] 10) When a historical search term carrying seasonal attribute information is detected, the historical search term carrying seasonal attribute information is obtained, and a historical search period of the historical search term is obtained.

[0138] 11) According to the search times of the historical search term on the current application platform, a first adjustment parameter corresponding to the search times is determined.

[0139] 12) According to the historical search hotness data of the historical search term in the historical search period, the search times of the historical search term on the current application platform, and the first adjustment parameter, a recommendation index corresponding to the historical search term is determined.

[0140] 13) According to the recommendation index, a search hotness value of each historical search term is determined.

[0141] 14) The historical search terms are sorted according to the search hotness values, and a corresponding search term list is generated.

[0142] 15) According to the seasonal attribute information, the corresponding search term list is adjusted in real time in a display time period corresponding to the historical search period, and is displayed.

[0143] In the above search recommendation term generation method, by obtaining the historical search term carrying the seasonal attribute information and obtaining the historical search period of the historical search term, the recommendation index corresponding to the historical search term can be determined according to the historical search hotness data of the historical search term in the historical search period and the search times of the historical search term on the current application platform. Further, according to the recommendation index and the seasonal attribute information, the generated search recommendation term list is adjusted and displayed. It is realized that for the historical search term with seasonal attribute information, the generated search term recommendation list can be flexibly adjusted according to the recommendation index and the seasonal attribute information, so as to realize the positive recommendation and guidance of users to search and purchase products at the appropriate time point, avoid invalid recommendation, and improve the correlation between the recommendation terms displayed by the application platform and the users, and the guidance effect and recommendation effect brought by the recommendation terms.

[0144] It should be understood that although the above embodiments involve the steps in each flowchart being shown in sequence according to the arrows, the steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, the steps are not strictly limited in sequence, and the steps can be executed in other sequences. Moreover, at least some of the steps in each flowchart involved in the above embodiments can include multiple steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution sequence of the steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least part of other steps or steps or stages in other steps.

[0145] In one embodiment, as shown in FIG. 15, a search recommendation word generation device is provided, which can be a software module or a hardware module, or a combination of the two as part of a computer device, and specifically includes: a historical search word acquisition module 1502, a recommendation index determination module 1504, and a search recommendation word list generation module 1506, wherein: Figure 15

[0146] The historical search word acquisition module 1502 is configured to acquire historical search words carrying seasonal attribute information, and acquire the historical search period of the historical search words.

[0147] The recommendation index determination module 1504 is configured to determine the recommendation index corresponding to the historical search word according to the historical search heat data of the historical search word in the historical search period and the search times of the historical search word on the current application platform.

[0148] The search recommendation word list generation module 1506 is configured to adjust and display the generated search recommendation word list according to the recommendation index and the seasonal attribute information.

[0149] In this embodiment, by acquiring historical search words carrying seasonal attribute information and acquiring the historical search period of the historical search words, the recommendation index corresponding to the historical search word can be determined according to the historical search heat data of the historical search word in the historical search period and the search times of the historical search word on the current application platform. Further, the generated search recommendation word list is adjusted and displayed according to the recommendation index and the seasonal attribute information. It is realized that for historical search words with seasonal attribute information, the generated search word recommendation list can be flexibly adjusted according to the recommendation index and the seasonal attribute information, so as to realize positive recommendation and guide users to search and purchase products at the right time, avoid invalid recommendation, and improve the association between the recommendation words displayed by the application platform and the users, and the guiding effect and recommendation effect brought by the recommendation words.

[0150] ​In one embodiment, the recommendation index determination module is further configured to:

[0151] determine a first adjustment parameter corresponding to the search frequency according to the search frequency of the historical search word on the current application platform; and determine the recommendation index corresponding to the historical search word according to the historical search word's whole-network historical search heat data in the historical search period, the search frequency of the historical search word on the current application platform, and the first adjustment parameter.

[0152] In one embodiment, a search recommendation word generation device is provided, which further comprises:

[0153] a historical search data collection module configured to collect historical search data of the current application platform and a whole-network search engine in a first preset time period;

[0154] a historical search word extraction module configured to traverse the historical search data and extract historical search words associated with the time period;

[0155] a seasonal attribute analysis module configured to perform seasonal attribute analysis on the historical search words associated with the time period to determine whether the historical search words carry seasonal attribute information.

[0156] In one embodiment, the seasonal attribute analysis module is further configured to:

[0157] obtain the time points of occurrence of the historical search words, and perform search period analysis based on the time points of occurrence of the historical search words to generate corresponding historical search periods; extract the whole-network search frequencies of the historical search words in the corresponding historical search periods; determine the current average heat index of the historical search words based on the whole-network search frequencies; and perform seasonal attribute analysis on the historical search words according to the current average heat index and a preset heat threshold to determine whether the historical search words carry seasonal attribute information.

[0158] In one embodiment, the seasonal attribute analysis module is further configured to:

[0159] determine a heat index vertex greater than the preset heat threshold according to the current average heat index and the corresponding preset heat threshold; obtain the time of occurrence of the heat index vertex in previous years, and determine the time intervals between the occurrence times; and perform seasonal attribute analysis on the historical search words according to the time intervals between the occurrence times and the corresponding preset interval time threshold to determine whether the historical search words carry seasonal attribute information.

[0160] In one embodiment, the seasonal attribute analysis module is further configured to:

[0161] The mobile average heat data corresponding to the historical search word is generated by performing mobile average processing on the search times of the whole network. The annual average heat data of the historical search word is obtained, and the current average heat index of the historical search word is generated according to the mobile average heat data and the annual average heat data.

[0162] In one embodiment, the search recommendation word list generation module is further configured to:

[0163] According to the recommendation index, the search heat value of each historical search word is determined. The search word list is generated by sorting according to the search heat value of each historical search word. The search word list corresponding to the display time period corresponding to the historical search period is generated in real time according to the seasonal attribute information, and is displayed.

[0164] In one embodiment, a search recommendation word generation device is provided, and the device further comprises:

[0165] The real-time hot keyword acquisition module is configured to traverse each historical search data to obtain real-time hot keywords.

[0166] The product keyword extraction module is configured to perform named entity recognition processing on each real-time hot keyword to extract product keywords matched with product attributes.

[0167] The historical search heat acquisition module is configured to obtain the historical search heat of each product keyword in a second preset time period and the search heat growth multiple.

[0168] The second adjustment parameter determination module is configured to determine the corresponding second adjustment parameter according to the search heat growth multiple.

[0169] The product keyword recommendation index determination module is configured to determine the recommendation index of the product keyword according to the historical search heat, the search heat growth multiple, and the second adjustment parameter corresponding to the search heat growth multiple.

[0170] The real-time recommendation word list generation module is configured to generate and display the corresponding real-time recommendation word list based on the recommendation index corresponding to each product keyword.

[0171] The specific limitations of the search recommendation word generation device can be referred to the limitations of the search recommendation word generation method in the foregoing, and will not be repeated here. Each module in the search recommendation word generation device described above can be realized by software, hardware, and combinations thereof, in whole or in part. The above-mentioned modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so as to be called and executed by the processor to perform the operations corresponding to the above-mentioned modules.

[0172] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 16 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores historical search terms, historical search cycles, historical search popularity data across the entire network, search counts, recommendation indices, seasonal attribute information, and a list of recommended search terms. The network interface is used for communication with external terminals via a network connection. When executed by the processor, the computer program implements a method for generating recommended search terms.

[0173] Those skilled in the art will understand that Figure 16 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0174] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0175] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0176] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the computer device to perform the steps in the above method embodiments.

[0177] 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. The above-mentioned computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, the processes of the above-mentioned embodiments of each method can be included. Any reference to memory, database or other medium used in each embodiment provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The database involved in each embodiment provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in each embodiment provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0178] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of each technical feature in the above embodiments are not described, however, as long as the combination of the technical features does not exist, it should be considered as the scope of the present application.

[0179] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of the present application. Therefore, the scope of protection of the patent of the present application should be subject to the appended claims.

Claims

1. A search recommendation word generation method characterized by comprising: The method comprises: acquiring historical search words carrying seasonal attribute information and acquiring a historical search period of the historical search words; determining a recommendation index corresponding to the historical search words according to historical search heat data of the historical search words in the historical search period and a search frequency of the historical search words on a current application platform; adjusting a generated search recommendation word list according to the recommendation index and the seasonal attribute information and displaying the search recommendation word list; The method further comprises: traversing each historical search data to acquire real-time hot keywords; performing named entity recognition processing on each real-time hot keyword to extract product keywords matching product attributes; acquiring historical search heat of each product keyword in a second preset time period and a search heat growth multiple; determining a corresponding second adjustment parameter according to the search heat growth multiple; determining a recommendation index of the product keyword according to the historical search heat, the search heat growth multiple, and the second adjustment parameter corresponding to the search heat growth multiple; and generating a corresponding real-time recommendation word list based on the recommendation index corresponding to each product keyword and displaying the real-time recommendation word list.

2. The method of claim 1, wherein, The determination of the recommendation index corresponding to the historical search words according to the historical search heat data of the historical search words in the historical search period and the search frequency of the historical search words on the current application platform comprises: determining a first adjustment parameter corresponding to the search frequency according to the search frequency of the historical search words on the current application platform; determining the recommendation index corresponding to the historical search words according to the historical search heat data of the historical search words in the historical search period, the search frequency of the historical search words on the current application platform, and the first adjustment parameter.

3. The method of claim 1, wherein, Before the acquisition of the historical search words carrying seasonal attribute information and the acquisition of the historical search period of the historical search words, the method further comprises: collecting historical search data of a current application platform and a global search engine in a first preset time period; traversing each historical search data to extract historical search words associated with a time period; performing seasonal attribute analysis on the historical search words associated with the time period to determine whether each historical search word carries seasonal attribute information.

4. The method of claim 3, wherein, The seasonal attribute analysis on the historical search words associated with the time period to determine whether each historical search word carries seasonal attribute information comprises: acquiring an occurrence time point of each historical search word and performing search period analysis based on the occurrence time point of each historical search word to generate a corresponding historical search period; extracting a global search frequency of each historical search word in the corresponding historical search period; determining a current average heat index corresponding to each historical search word based on the global search frequency; performing seasonal attribute analysis on each historical search word according to the current average heat index and a preset heat threshold to determine whether each historical search word carries seasonal attribute information.

5. The method of claim 4, wherein, The method comprises the following steps: According to the current average heat index and the preset heat threshold, a heat index vertex greater than the preset heat threshold is determined; The time of appearance of the heat index vertex in the past years is obtained, and the time interval between each of the time of appearance is determined; According to the time interval between the time of appearance and the corresponding preset interval threshold, the seasonal attribute analysis of the historical search words is performed to determine whether each of the historical search words carries seasonal attribute information.

6. The method according to any one of claims 1 to 3, characterized in that, The method comprises the following steps: According to the recommendation index, the search heat value of each of the historical search words is determined; According to the search heat value of each of the historical search words, a corresponding search word list is generated by sorting; According to the seasonal attribute information, the corresponding search word list is generated in real time in the display time period corresponding to the historical search period and is displayed.

7. The method of claim 4, wherein, The method comprises the following steps: According to the network-wide search times, the current average heat index corresponding to the historical search words is determined; According to the network-wide search times, the moving average heat data corresponding to the historical search words is generated; 8. A search recommendation word generating apparatus characterized by comprising: The annual average heat data of the historical search words is obtained, and the current average heat index of the historical search words is generated according to the moving average heat data and the annual average heat data. The device comprises: A historical search word acquisition module is configured to acquire historical search words carrying seasonal attribute information and the historical search period of the historical search words; A recommendation index determination module is configured to determine the recommendation index corresponding to the historical search words according to the network-wide historical search heat data of the historical search words in the historical search period and the search times of the historical search words on the current application platform; A search recommendation word list generation module is configured to adjust the generated search recommendation word list according to the recommendation index and the seasonal attribute information and display the search recommendation word list; The device further comprises: a real-time hot keyword acquisition module configured to traverse each of the historical search data to acquire real-time hot keywords; a product keyword extraction module configured to perform named entity recognition processing on each of the real-time hot keywords to extract product keywords matched with product attributes; a historical search heat acquisition module configured to acquire the historical search heat of each of the product keywords in a second preset time period and the search heat growth multiple; a second adjustment parameter determination module configured to determine the corresponding second adjustment parameter according to the search heat growth multiple; a product keyword recommendation index determination module configured to determine the recommendation index of the product keywords according to the historical search heat, the search heat growth multiple, and the second adjustment parameter corresponding to the search heat growth multiple; and a real-time recommendation word list generation module configured to generate a corresponding real-time recommendation word list based on the recommendation index corresponding to each of the product keywords and display the real-time recommendation word list.

9. The apparatus of claim 8, wherein, The recommendation index determination module is further configured to: determine a first adjustment parameter corresponding to the search frequency according to the search frequency of the historical search word on the current application platform; and determine a recommendation index corresponding to the historical search word according to the historical search word's whole-network historical search popularity data in the historical search period, the search frequency of the historical search word on the current application platform, and the first adjustment parameter.

10. The apparatus of claim 8, wherein, The device further comprises: a historical search data collection module configured to collect historical search data of a current application platform and a whole-network search engine in a first preset time period; a historical search word extraction module configured to traverse the historical search data and extract historical search words associated with a time period; a seasonal attribute analysis module configured to perform seasonal attribute analysis on the historical search words associated with the time period and determine whether each of the historical search words carries seasonal attribute information.

11. The apparatus of claim 10, wherein, The seasonal attribute analysis module is further configured to: obtain time points at which each of the historical search words appears, and perform search period analysis based on the time points at which each of the historical search words appears to generate corresponding historical search periods; and extract whole-network search frequencies of each of the historical search words in the corresponding historical search periods. determine a current average popularity index corresponding to each of the historical search words based on the whole-network search frequencies; and perform seasonal attribute analysis on each of the historical search words according to the current average popularity index and a preset popularity threshold to determine whether each of the historical search words carries seasonal attribute information.

12. The apparatus of claim 11, wherein, The seasonal attribute analysis module is further configured to: determine a popularity index vertex greater than the preset popularity threshold according to the current average popularity index and the corresponding preset popularity threshold; obtain times at which the popularity index vertex appears in previous years, and determine time intervals between each of the times; perform seasonal attribute analysis on each of the historical search words according to the time intervals between the times and a corresponding preset interval time threshold to determine whether each of the historical search words carries seasonal attribute information.

13. The apparatus of any one of claims 8 to 10, wherein, The search recommendation word list generation module is further configured to: determine search popularity values of each of the historical search words according to the recommendation indexes; sort the search word list according to the search popularity values of each of the historical search words to generate a corresponding search word list; and adjust the corresponding search word list in real time in a display time period corresponding to the historical search period according to the seasonal attribute information and display the search word list.

14. The apparatus of claim 11, wherein, The seasonal attribute analysis module is further configured to: perform moving average processing on the whole-network search frequencies to generate moving average popularity data corresponding to each of the historical search words; obtain annual average popularity data of each of the historical search words; and generate a current average popularity index of each of the historical search words according to the moving average popularity data and the annual average popularity data.

15. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor, when executing the computer program, implements the steps of the method of any one of claims 1 to 7.

16. A computer readable storage medium storing a computer program, wherein the computer program comprises program instructions configured to cause a processor to perform the method according to any one of claims 1 to 15. The computer program, when executed by the processor, implements the steps of the method of any one of claims 1 to 7.

17. A computer program product comprising a computer program, characterized in that, The computer program, when executed by the processor, implements the steps of the method of any one of claims 1 to 7.

Images