A kind of Method of Commodity Recommendation, recommendation apparatus and refrigerator
Technical field
The present invention relates to data screening technical field, specifically for, be a kind of Method of Commodity Recommendation, recommendation apparatus and
Refrigerator.
Background technology
Along with the development of the Internet, people are accustomed to shopping online already, but, how the product listed on the net is selected
Selecting, people are created puzzlement, for solving this problem, commercial product recommending is particularly important, and conventional recommendation method has as follows
Two kinds.
(1) manual configuration: configure items list on backstage by operation personnel, then, the artificial selection of operation personnel
Under, manual configuration needs the product recommended.The method is the starting stage of store intermediate item commercial product recommending--pass through operation personnel
Manual configuration mode carries out commercial product recommending, but, this method subjectivity, randomness are too strong, do not have unified commercial product recommending mark
Standard, implementation capacity are relatively low, need substantial amounts of operation personnel, expend substantial amounts of manpower, and, immediately in the feelings that operation personnel is very many
Under condition, also it is difficult to when commodity are too much be safeguarded by backstage, more cannot meet the needs of all users, it addition, this kind of mode
Do not indicate item property, be difficult to safeguard the similarity of commodity, the recommendation effect extreme difference of commodity.
(2) fixed index dimension is recommended: the above-mentioned manual configuration method recommendation effect for solving is poor, manpower waste is serious
Problem, someone have employed the recommendation method of fixed index dimension, as carried out recommended products according to time to market (TTM).But, this
Method recommend commodity Indexes Comparison the most single, reference fix, it is recommended that commodity number and index ratio cannot join flexibly
Putting, thus be only applicable to the user of very fraction, major part user is likely to and is not concerned with These parameters dimension, therefore, this
Method of Commodity Recommendation narrow application range, recommendation effect are the most very poor.
Therefore, it is thus achieved that a kind of commercial product recommending is effective, flexibly configurable commodity number and index ratio, be applicable to major part
The Method of Commodity Recommendation of user becomes those skilled in the art's technical problem urgently to be resolved hurrily and the emphasis of research.
Summary of the invention
For solving that the human input that existing goods recommends method to exist is big, recommendation effect is poor, it is suitable for that user scope is narrow etc. asks
Topic, the invention provides a kind of Method of Commodity Recommendation, recommendation apparatus and refrigerator, can complete the work of intelligentized commercial product recommending, push away
Recommending effect preferable, human input amount is greatly reduced, and is suitable for user scope wide.
For realizing above-mentioned technical purpose, the invention discloses a kind of Method of Commodity Recommendation, the method comprises the steps:
Step 1, obtains the commodity data of active user location;
Step 2, screens above-mentioned commodity data by least one mode in table in red-letter day, History Order, Sales Volume of Commodity;
Step 3, utilizes the commodity data after screening to obtain the merchandise news needing to recommend.
By means of the commodity data in this area, the Method of Commodity Recommendation that the present invention provides is being not known by the letter of active user
Reasonably recommending under breath, the method breaches traditional way of recommendation, innovatively by table in red-letter day, History Order, commodity
Sales volume angularly provides commodity recommendation information for active user, the method recommendation effect is good, human input is little, be suitable for push away on a large scale
Wide use.
Further, in step 1, described commodity data includes label data, evaluating data, commodity special price data, sales volume
In data at least one.
Further, in step 2, obtain the red-letter day nearest with current time, calculate current date and described date in red-letter day
The absolute value of difference, it is judged that whether this absolute value is less than threshold value T:
If this absolute value less than threshold value T, then filters out in the commodity data from step 1 has respective labels in described red-letter day
Commodity data;
If this absolute value is more than or equal to threshold value T, then screened by least one mode in History Order, Sales Volume of Commodity
Above-mentioned commodity data.
Further, in step 2, if this absolute value is less than threshold value T, is had by the screening of comment situation and be correlated with described red-letter day
The commodity data of label, win the favorable judgment forward commodity data.
Further, in step 2, by whether the mode for bargain goods screens the commodity data that favorable comment is forward;Step 3
In, the bargain goods obtained screens according to shelf life.
Further, in step 2, if this absolute value is more than or equal to threshold value T, it is judged that in active user or this area its
Whether he user has a History Order:
If there being History Order, then screen commodity number according to the label data in user's History Order and/or goods browse amount
According to;
If without History Order, then according to sales volume data screening commodity data.
Further, in step 2, label data includes at least one in mouthfeel, effect, nutrition.
Further, in step 2, described threshold value T is 7 days.
Further, in step 3, real-time tracking recommendation results, the purchase for active user selects again to screen commodity
Data, adjustment need the product recommended.
Further, in step 3, the described merchandise news needing to recommend is shown on refrigerator display screen.
Another goal of the invention of the present invention is to provide a kind of device for recommending the commodity, and this device includes display screen and backstage
Processing equipment, described background process equipment includes:
Acquisition module, this module is for obtaining the commodity data of active user location;
Screening module, this module screens above-mentioned commodity by least one mode in table in red-letter day, History Order, Sales Volume of Commodity
Data;
Output module, this module utilizes the commodity data after screening to obtain the merchandise news needing to recommend, and by these commodity
Information is showed on display screen.
The device for recommending the commodity disclosed by the invention demonstrates, by display screen, the merchandise news that background process equipment calculates,
Background process equipment can be located on server, it is possible to is installed on corresponding carrier, on the products such as refrigerator.
The present invention also has a goal of the invention to be to provide a kind of refrigerator, the device for recommending the commodity that refrigerator is above-mentioned, display screen
It is fixed on refrigerator.
The present invention proposes a kind of refrigerator with commercial product recommending function, and this refrigerator can be new user in the cold start mode
Reasonably Recommendations, it is recommended that effective, human input is few, applied widely.
The invention have the benefit that the Method of Commodity Recommendation of the present invention has dynamic proportion screening function, by referring to more
Mark dimension, label also combine dynamic proportion and screen, and flexibly configurable commodity number and index ratio are suitable for almost all
New user and old user, the effect of commercial product recommending is the best.
The present invention also has the motility of raising commercial product recommending rule configuration, alleviates the manual intervention of operation personnel, intelligence
Adjust and download step-length, download the advantages such as data with optimal state.
It addition, at method design aspect, the present invention can be designed by clear data storehouse completely, it is not necessary to combine development language real
Existing, flexibly configurable commercial product recommending ratio and condition, commercial product recommending data periodically automatically generate, and the present invention can be completely achieved intelligence
Change, it is not necessary to user's manual operation.
Accompanying drawing explanation
Fig. 1 is Method of Commodity Recommendation flow chart of the present invention.
Fig. 2 is the Method of Commodity Recommendation of the present invention flow chart when implementing according to concrete screening conditions.
Fig. 3 is the data access flow process figure of Method of Commodity Recommendation of the present invention.
Detailed description of the invention
Below in conjunction with Figure of description, the present invention is carried out detailed explanation and explanation.
The invention discloses a kind of Method of Commodity Recommendation, the method includes index dimension combing, metadata schema design, stream
Journey is drawn, metadata generates, commercial product recommending result set extracts, recommend tracking.
Wherein, index dimension combing segment includes recommending the fast-selling degree according to commodity, positive rating, special price situation, label to divide
Class, History Order situation;Metadata schema design mainly includes region, label, the design of labeling backstage table;Flow process is drawn
Logical judgment detailed process including commercial product recommending is shown with the form of instrument;Metadata generate be city distribution data, label,
The metadata such as labeling, each dimension withdrawal ratio are safeguarded;It is to combine according to the thinking of flow chart that commercial product recommending result set extracts
Self-defining accounting generates the items list of regional;Wherein recommending tracking is to be tracked recommendation effect, analyze recommendation
Effect is to adjust Generalization bounds in time.
As shown in Figure 1, 2, the method specifically includes following steps:
Step 1, obtains the commodity data of active user location;Commodity data is interpreted as the data relevant with commodity,
In the present embodiment, commodity data include commodity basic data, label data, evaluating data, commodity special price data, sales volume data,
User's order data, geographic classification data, labeling data, table in red-letter day, recommend at least one in the data such as dynamic parameter.
Step 2, screens above-mentioned commodity data by least one mode in table in red-letter day, History Order, Sales Volume of Commodity;Can press
Screen according to table in red-letter day, History Order, Sales Volume of Commodity order, obtain the red-letter day nearest with current time, calculate current date
Absolute value with date differences in described red-letter day, it is judged that whether this absolute value is less than threshold value T, under the technology of the present invention is enlightened, should
The big I of threshold value T carries out the rationally selection of wisdom according to time situation, and preferred threshold value T of the present invention is 7 days.
If this absolute value less than threshold value T, then filters out in the commodity data from step 1 has respective labels in described red-letter day
Commodity data, illustrate that the commodity that these commodity datas are corresponding are labeled with label in red-letter day in absolute value natural law, pass through label list
Obtain being under the jurisdiction of this area's property in red-letter day product, in the present embodiment, by the product that red-letter day, table screened with the commodity association that can provide and deliver
Account for the 40% of whole Recommendations;Then be there is by the screening of comment situation the commodity data of respective labels in described red-letter day, obtain
The commodity data that favorable comment is forward, in the present embodiment, the forward commodity of favorable comment account for the 20% of whole Recommendations, on above-mentioned basis
On, by whether the mode for bargain goods screens the commodity data that favorable comment is forward, in the present embodiment, bargain goods accounts for whole pushing away
Recommend the 40% of commodity.
If this absolute value is more than or equal to threshold value T, then screened by least one mode in History Order, Sales Volume of Commodity
Above-mentioned commodity data.Specifically, it is judged that in active user or this area, whether other users have History Order: if there being history to order
Single, then screen commodity data, in the present embodiment, label according to the label data in user's History Order and/or goods browse amount
Data include at least one in mouthfeel, effect, nutrition, and the commodity after screening account for the 40% of whole Recommendations;Further, this reality
Executing example and can screen a month interior History Order further, the commodity after screening account for the 20% of whole Recommendations;If without going through
History order, then according to sales volume data screening commodity data, fast-selling commodity account for the 20% of whole Recommendations.Above-mentioned recommended case
In, if total accounting is unsatisfactory for 100%, then according to adjusting recommendation ratio or random way of recommendation solution.
Step 3, for the bargain goods obtained in step 2, can screen according to shelf life.Utilize the commodity number after screening
According to obtaining the merchandise news needing to recommend, it is achieved be user's Recommendations under conditions of without any information of active user.For complete
Kind recommendation method, real-time tracking recommendation results of the present invention, this recommendation results refers to whether user buys some commodity or browse meaning
To commodity etc., select again to screen commodity data, adjust the product needing to recommend for buying of active user, use according to current
Family History Order analyzes the buying habit of active user.The product that the present invention recommends is all can to provide and deliver in active user location
, the commodity as recommended are can to provide and deliver in city, refrigerator place.
As Fig. 3 provides above-mentioned Method of Commodity Recommendation data access flow process figure, current locale refers to urban addresses or refrigerator longitude and latitude
Degree information, it may be judged whether incoming city id, if it is by this geographical position, as defaulted to certain city, the business that display is recommended
Product information, if it is not, then the merchandise news of display general recommendations;After incoming city id, it may be judged whether incoming user id, if had
Then gone the merchandise news of coupling recommendation by user id, returning a user id is not empty recommendation results collection, without then
The merchandise news that display is recommended;Then judge whether that incoming reception bar number limits, if it is, go coupling by the bar number limited
The merchandise news recommended, without the merchandise news then showing acquiescence;Default value is 10.The present invention is called by foreground and pushes away
Recommend data access interface, coordinate the page to be shown. acquiescence backstage periodic refreshing every day Recommendations.
It should be strongly noted that this method is combined by the present invention with refrigerator, the result of the commodity recommended is needed to show
On refrigerator display screen, reach to make full use of the purpose of refrigerator large-screen display.Above-mentioned screening mode dynamically configures, such as,
Every day filter information judged and update the information of Recommendations.
For implementing above-mentioned Method of Commodity Recommendation, the invention also discloses a kind of device for recommending the commodity, this device includes showing
Display screen and background process equipment, background process equipment includes:
Acquisition module, this module, for obtaining the commodity data of active user location, plays the effect of data collection,
Commodity data includes at least one in label data, evaluating data, commodity special price data, sales volume data.Acquisition module is used for holding
The step 1 of the above-mentioned Method of Commodity Recommendation of row.
Screening module, this module screens above-mentioned commodity by least one mode in table in red-letter day, History Order, Sales Volume of Commodity
Data, can realize by the way of concrete hardware device or software and hardware combining, and screening module is used for performing above-mentioned commercial product recommending
The step 2 of method.
Output module, this module utilizes the commodity data after screening to obtain corresponding with commodity data, to need recommendation business
Product information, and this merchandise news is showed on display screen.Output module is for performing the step 3 of above-mentioned Method of Commodity Recommendation.
The present invention also discloses a kind of refrigerator, can improve, make ice on the refrigerator of current refrigerator or new production
There is on case the device for recommending the commodity, and display screen is fixed on refrigerator.Specifically, by means of Internet of Things, the present invention can make visitor
The functions such as commodity purchasing, video-see, menu reference, food materials configuration and refrigerator are arranged are experienced by the display screen on refrigerator in family,
The present invention solves under refrigerator cold start-up mode, do not understand user trajectory in the case of carry out reasonable commercial product recommending problem for client,
" cold start-up mode " can be regarded as refrigerator and do not record any information of active user.The invention provides one and there is commercial product recommending
The refrigerator of function, it is recommended that select according to index dimension brush, according to label, labeling screening commodity, sieves according to order correlation tag
Select commodity, to index ratio, recommend sum to realize dynamically adjustment.
The foregoing is only presently preferred embodiments of the present invention, not in order to limit the present invention, all in essence of the present invention
Any amendment, equivalent and the simple modifications etc. made in content, should be included within the scope of the present invention.