KR102346429B1 - Fruit recommendation system and method based on artificial intelligence curation - Google Patents

Abstract

The present invention relates to a fruit recommendation system and method based on AI curation, and is for automatically recommending fruit composition by AI curation using big data on fruit and customers' purchasing patterns. The AI curation-based fruit recommendation system according to the present invention includes a fruit recommendation server capable of communicating fruit purchase information with a producer terminal and a consumer terminal, and recommending fruit suitable for a user's transaction pattern, wherein the fruit recommendation server is, an information collection unit that collects information from the producer terminal and the consumer terminal, a database unit in which the collected information is stored, and extracts a consumer's transaction pattern through big data analysis of the information stored in the database unit to obtain product information and It includes a big data analysis unit that analyzes in connection with user information and a fruit selection unit that selects fruits to be delivered customized to the consumer according to the information analyzed by the big data analysis unit.

Description

AI Curation-based Fruit Recommendation System and Method {FRUIT RECOMMENDATION SYSTEM AND METHOD BASED ON ARTIFICIAL INTELLIGENCE CURATION}

The present invention relates to an AI curation-based fruit recommendation system and method, and more particularly, to a system and method for automatically recommending fruit composition by AI curation using big data on fruit and customer purchasing patterns. will be.

The background of the present invention is that, in the existing fruit purchase system, the fruit composition of seasonal subscription fruits is fixed according to the order amount, and the fruit packer arbitrarily configures and delivers seasonal fruits according to the order unit price. However, it is difficult to apply for fruit selection and fruit subscription services to suit the diverse tastes of consumers. Furthermore, the problems of the existing system are that customers may receive fruits that do not fit or want to be left behind when purchasing or subscribing fruits, and that even if the customer wants to change the composition of fruits arbitrarily, it is difficult because the customer has little information about fruits. The point is that in the case of fruits, since the seasonal fruits are constantly changing, there is no choice but to choose the fruit packer arbitrarily.

Republic of Korea Patent Publication No. 10-2016-0106445 (published on September 12, 2016)

An object of the present invention to solve the above problems is to provide a fruit recommendation system and method based on AI curation that can improve customer satisfaction by delivering fruits that are helpful for customer preference and health care through AI curation will do

Another object of the present invention is to provide a fruit recommendation system and method based on AI curation that can reduce fruit inventory through quantity prediction and ordering according to fruit composition through AI curation.

Another object of the present invention is to provide a fruit recommendation system and method based on AI curation that can reduce the amount of leftover fruit by receiving a customer's preferred fruit.

Another object of the present invention is to provide a fruit recommendation system and method based on AI curation that can improve the work efficiency of fruit packing staff by packaging and shipping based on fruit type and number information provided by AI curation.

In order to achieve the above object, the AI curation-based fruit recommendation system according to an embodiment of the present invention enables communication about fruit purchase information with a producer terminal and a consumer terminal, and recommends fruits suitable for the user's transaction pattern. A fruit recommendation server, wherein the fruit recommendation server performs big data analysis of an information collection unit that collects information from the producer terminal and the consumer terminal, a database unit in which the collected information is stored, and the information stored in the database unit It includes a big data analysis unit that extracts the consumer's transaction pattern and analyzes it in connection with product information and user information, and a fruit selection unit that selects fruits to be delivered customized to the consumer according to the information analyzed by the big data analysis unit.

The information collection unit collects efficacy, characteristics, and price information of each of the various types of fruits from the producer terminal, and collects transaction information and consumer information from the consumer terminal.

The database unit, from the user terminal, a fruit efficacy database unit in which anticancer, antidepressant and antioxidant information of the fruit is stored, respectively, a fruit characteristic database unit in which taste, color, nutrition and shipping information of the fruit are stored, respectively; and a fruit price database in which information about the fruit packaging unit of the fruit, the average weight per piece, the purchase price per unit weight, and the selling price per unit weight are respectively stored, and from the consumer terminal, the consumer's purchase frequency, purchase amount, and purchase satisfaction information It includes a fruit transaction database unit that is stored respectively, a consumer database unit that stores the consumer's search history, shopping cart information, gender, age, taste, preference, environment-friendly and allergy information, respectively, and a purchase purpose database unit that stores the purchase purpose of the consumer do.

The fruit recommendation server receives a purchase price from the consumer terminal, and the fruit selection unit selects the fruit according to the received purchase price and the consumer's purchase pattern.

In order to achieve the above object, the AI curation-based fruit recommendation method according to an embodiment of the present invention includes the steps of receiving a purchase price from a consumer terminal, delivering the purchase amount to the producer terminal, and from the producer terminal The steps of receiving information about circulating fruit, analyzing the information collected from the consumer terminal to extract the consumer's transaction pattern, and selecting the fruit to be delivered to the consumer in a customized way according to the extracted consumer's transaction pattern and the purchase amount including the steps of

The present invention can provide a fruit recommendation system and method based on AI curation that can improve customer satisfaction by delivering fruits that are helpful for customer preference and health care through AI curation.

The present invention can provide a fruit recommendation system and method based on AI curation that can reduce fruit inventory through quantity prediction ordering according to fruit composition through AI curation.

The present invention can provide a fruit recommendation system and method based on AI curation that can reduce the amount of fruit left over by receiving the fruit preferred by the customer.

The present invention can provide a fruit recommendation system and method based on AI curation that can improve the work efficiency of fruit packing staff by packaging and shipping based on the fruit type and number information provided by AI curation.

1 is a block diagram schematically showing the configuration included in the AI curation-based fruit recommendation system according to the present invention.
2 is a block diagram schematically showing each configuration of the database unit according to the present invention.
3 is a flowchart showing a flow according to the AI curation-based fruit recommendation method according to the present invention.

Hereinafter, some embodiments of the present invention will be described in detail with reference to exemplary drawings. In adding reference numerals to components in each drawing, it should be noted that the same components are given the same reference numerals as much as possible even though they are indicated in different drawings.

In the description of the embodiment of the present invention, if it is determined that a detailed description of a related known configuration or function interferes with the understanding of the embodiment of the present invention, the detailed description thereof will be omitted.

In addition, in describing the components of the embodiment of the present invention, terms such as first, second, A, B, (a), (b), etc. may be used. These terms are only for distinguishing the elements from other elements, and the essence, order, or order of the elements are not limited by the terms.

Hereinafter, the present invention will be described in more detail with reference to the accompanying drawings.

1 is a block diagram schematically showing the configuration included in the AI curation-based fruit recommendation system according to the present invention.

Referring to FIG. 1 , the AI curation-based fruit recommendation system according to the present invention enables communication about fruit purchase information with the producer terminal 20 and the consumer terminal 10, and selects fruits suitable for the user's transaction pattern. It includes a recommended fruit recommendation server (100).

The fruit recommendation server 100 may further include a communication unit (not shown) capable of two-way communication with the producer terminal 20 and the consumer terminal 10 . The fruit recommendation server 100 is connected to the producer terminal 20 and the consumer terminal 10 through a communication network, and recommends suitable fruits according to the order time or order amount by deriving a fruit transaction pattern from the consumer terminal 10 . can do it

The fruit recommendation server 100 includes an information collection unit 110 for collecting information from the producer terminal 20 and the consumer terminal 10 , a database unit 120 for storing information collected by information collection, and a database unit 120 . ) to consumers by the information analyzed by the big data analysis unit 130 and the big data analysis unit 130, which extracts the consumer's transaction pattern through big data analysis and analyzes the information stored in the fruit information and consumer information. It further includes a fruit selection unit 140 for selecting fruits to be delivered in a customized manner.

The information collection unit 110 may collect information on the efficacy, characteristics, and price of each of the various types of fruits from the producer terminal 20 . The producer terminal 20 may input information about each fruit toward the fruit recommendation server 100 . The information collection unit 110 may collect each information on the fruit that can be distributed for each season or each time from the producer terminal 20 , and may also collect information on the supply and demand situation and inventory of fruit for each season. The information collected by the information collection unit 110 may be respectively stored in the database unit 120 to be described later. On the other hand, the information collected by the information collection unit 110 is not limited to the information input from the producer terminal 20, and information can also be collected from the Internet web that can be accessed through an online communication network.

In more detail, the efficacy information collected by the information collection unit 110 includes information helpful for anticancer, antidepressant, and antioxidant properties of each fruit, information helpful for digestive, circulatory, respiratory, and endocrine disturbance, Information about allergies may be collected. In addition, the information collection unit 110 may collect information on fruits that are helpful for emotions such as tension, expectation, anger, joy and sadness, in addition to anti-depression.

In relation to the anticancer information of the fruits collected by the information collection unit 110, for example, fruits effective in anticancer include tangerines, strawberries, grapes, tomatoes, and apples. A brief description of the anticancer effect of each fruit is as follows. For example, tangerine contains about 40 mg of vitamin C per 100 g, and in a study on Korean women, there was a study that vitamin C intake reduced the incidence of breast cancer. Grapes are rich in flavonoids. Studies have shown that certain flavonoids contained in grapes inhibit carcinogenesis such as colon cancer, breast cancer, skin cancer, liver cancer, and prostate cancer and reduce tumor size. Grape skins and seeds contain resveratrol, which has anticancer and powerful antioxidant properties. A research team at Rutgers University in the United States found that 70 women 55 years of age and older were given tomatoes daily for 10 weeks and then observed the hormone levels, and found that the secretion of adynecpotin increased. In general, when a woman's hormonal changes before and after menopause meets obesity, the risk of breast cancer increases. Apples contain pectin, a type of dietary fiber. Pectin is known to be very effective in preventing colon cancer. Studies have shown that the pectin in apples increases intestinal fatty acids, which protect against colorectal cancer. Apple juice is rich in polyphenols, and the polyphenols enhance the production of anticancer components during fermentation. In addition, with respect to endocrine disrupting, information on whether or not organic fruits are not grown with endocrine-disrupting chemicals such as pesticides or herbicides can be displayed to reduce exposure to endocrine-disrupting pesticides and fertilizers. Efficacy information collected by the information collection unit 110 may be stored in the fruit efficacy database unit 121 to be described later.

The characteristic information collected by the information collection unit 110 may include the taste, color, nutrition, and time of shipment of each fruit. The taste of each fruit may include sugar content, acidity, salinity, and the like. The color of each fruit may be distinguished by a color such as red, yellow, or purple. Nutrition of each fruit may include information such as nutrients and calories included on average per unit weight (preferably 100 g) of each fruit. The time of shipment of each fruit may be divided into spring, summer, autumn and winter, and the like, but is not limited thereto. For example, the delivery time of each fruit may be flexibly changed according to the production schedule of each fruit of the producer. That is, the values input from the producer terminal 20, such as month or day, are transmitted and may be changed in real time rather than as a season. In addition, information on other distribution status of each fruit may also be included. In addition, information on seasonal fruit supply and demand status and inventory may be included. The characteristic information collected by the information collection unit 110 may be stored in the fruit characteristic database unit 122, which will be described later.

The price information of each fruit collected by the information collection unit 110 may include a fruit packaging unit, an average weight per piece, a purchase price per unit weight (preferably 100 g), and a selling price per unit weight (preferably 100 g). . The fruit packaging unit may be a preset 3kg, 5kg, 10kg, 15kg, 20kg, etc., and may be arbitrarily set from the producer terminal 20 . The average weight per piece may be previously harvested and measured information. For example, 300 g of apples, 500 g of pears, 1 kg of melons, 80 g of plums, 100 g of kiwi, 10 g of cherries may be used, but is not limited thereto. The purchase price per unit weight is stored as a value inputted in real time from the producer terminal 20 . The selling price per unit weight may be calculated and determined based on the value input from the producer terminal 20 . The price information of each fruit collected by the information collection unit 110 may be stored in the fruit price database unit 123 which will be described later.

The information collection unit 110 may collect transaction information, consumer information, purchase purpose, and the like from the consumer terminal 10 .

The transaction information collected by the information collection unit 110 may include a purchase frequency, a purchase amount, and a purchase satisfaction level transacted from the consumer terminal 10 . The transaction information may be collected together with information directly input from the consumer terminal 10 in addition to information traded through the fruit recommendation server 100 according to the present invention. For example, transaction information that can be collected through a communication network excluding information stored in the fruit recommendation server 100 according to the present invention may also be collected. In this case, the identification number provided to the consumer terminal 10 through an identification number providing unit (not shown), which will be described later, may be used. The purchase frequency may be a past purchase frequency according to an identification number registered in the consumer terminal 10 . However, the present invention is not limited thereto, and for example, it may be an arbitrarily selected purchase frequency. Recently, a regular delivery service in which fruits are delivered regularly for a set period or on a specific date has been provided. In the case of an identification number for which regular delivery has been applied, the user may receive an input of a certain period, etc. that is stored when applying for regular delivery. The input period may replace the purchase frequency. The purchase amount may be an average value of the total amount of past purchases according to the identification number registered in the consumer terminal 10 and the purchase amount at one time of purchase. The purchase satisfaction may be the customer's satisfaction according to each item of pre-delivered fruit. The purchase satisfaction may be stored for each fruit, and may be separately collected and stored according to the efficacy and characteristics of each fruit. The transaction information collected by the information collection unit 110 may be stored in the fruit transaction database unit 124 which will be described later.

The consumer information collected by the information collection unit 110 may be information about gender, age group, fruit taste, preference, eco-friendly fruit, allergy, domestic fruit, and imported fruit that can be input from the consumer terminal 10 . In addition, the consumer information collected by the information collection unit 110 may include a search history of searching for fruits with the consumer terminal 10 and fruit information stored in a shopping cart for purchase.

Gender may be any one of male and female, and age may include youth, middle age, and old age. Fruit taste may be the taste, color, and nutrition of the fruit, and may be linked with fruit characteristic information collected by the information collection unit 110 from the producer terminal 20 . The preference may be a predetermined numerical value calculated according to the collected fruit taste, and may be calculated based on a weight value input when collecting each fruit taste information. For example, preference may receive input of each preference score for sugar, acidity, and salinity when receiving fruit taste from the consumer terminal 10, and may receive preference scores for low calorie in nutrition. A preference score may be determined as a value calculated by synthesizing each score. In addition, information indicating preference for eco-friendly fruits may be collected from the consumer terminal 10 . Eco-friendly fruits may refer to organic fruits that are not grown with chemicals such as pesticides or herbicides. The allergy information may be information input from the consumer terminal 10 . For consumers who are allergic to each fruit, the corresponding fruit may be excluded when the fruit is selected by the fruit selection unit 140 . Information on domestic fruits and imported fruits can be input from the consumer terminal 10 so that multiple selections are possible, and can be referenced when fruits are selected by the fruit selection unit 140 . For example, when the fruit selection unit 140 selects fruits, the selection of imported fruits may be excluded from the identification number that received only domestic fruits.

The information collection unit 110 may further collect information on the purchase purpose of the consumer. The consumer's purchase purpose may include diet use, use for test takers, use for job seekers, use of filial piety for parents, use of housewarming, use of gifts in return, use of visitation, health care use, and disease treatment use. The consumer's purchase purpose may be collected together when receiving the fruit purchase signal from the consumer terminal 10, and may be modified at any time.

2 is a block diagram schematically showing each configuration of the database unit 120 according to the present invention.

The AI curation-based fruit recommendation system according to the present invention includes a fruit recommendation server 100, and the fruit recommendation server 100 includes a database unit 120 in which information collected by the information collection unit 110 is stored. ) may be included. However, the present invention is not limited thereto, and, for example, it is possible to access the database unit 120 storing big data through a communication network. When accessing the database unit 120 through a communication network, the database unit 120 is not only a server having one specific IP (Internet Protocol) address, but also a plurality of servers connected through a communication network and having various different IP addresses. may be

Information respectively stored in the database unit 120 may be stored in the form of a mapping table. Each mapping table may be a relational database executed in relation to each other. A relational database is a collection of data items organized into a series of structured tables. Since the relational database is a known technology, a detailed description thereof will be omitted.

Referring to FIG. 2 , the database unit 120 according to the present invention includes a fruit efficacy database unit 121 , a fruit characteristic database unit 122 , and a fruit price database unit 123 in which information is collected and stored from a user terminal, respectively. and a fruit transaction database unit 124 , a consumer database unit 125 , and a purchase purpose database unit 126 in which information is collected and stored from the consumer terminal 10 , respectively.

In the fruit efficacy database unit 121, the various types of fruits collected by the information collection unit 110 each have information useful for anticancer, antidepressant, antioxidant, and emotions such as tension, expectation, anger, joy and sadness. information and the like may be stored. The information on anticancer, antidepressant and antioxidant stored in the fruit efficacy database unit 121 can be stored on the mapping table, and the existence or nonexistence of anticancer, antidepressant and antioxidant efficacy can be stored, and furthermore, a fruit recommendation server according to the present invention ( 100) If there is a research result that can be accessed through an external communication network, a weight value for each efficacy according to the research result may be included. A weight value for each efficacy may be computed complexly and a processed value may be stored. For example, the skin pigment of grapes is anthocyanin pigment and contains resveratrol among phenol components. In particular, resveratrol has anti-aging effects due to its antioxidant action in the body. It also has anticancer effects, as reported by the research results of inhibiting epithelial cell carcinoma in mice. Therefore, weights can be applied to fruits containing a lot of resveratrol per unit weight among various types of fruits in descending order by content. In addition, strawberries contain a large amount of physiologically active substances such as ellagic acid. As a plant phenolic acid, ellagic acid is known to have excellent anticancer properties, and is known to have antioxidant, AIDS virus suppression, and antimutagenic activity. Therefore, weights can be applied to fruits containing a large amount of ellagic acid per unit weight by arranging them in descending order by content. In addition to resveratrol and ellagic acid described as an example, the average value of the weight of fruits including ingredients effective for anticancer and the like may be stored in the fruit efficacy database unit 121 , respectively.

The fruit characteristic database unit 122 may store information about the taste, color, nutrition, shipping time, allergy, etc. of the various types of fruits collected by the information collection unit 110 . The fruit price database unit 123 may store information such as a fruit packaging unit, an average weight per piece, a purchase price per unit weight, and a selling price per unit weight of various types of fruits. The fruit transaction database unit 124 may store information about a consumer's purchase frequency, purchase amount, and purchase satisfaction, respectively. The consumer database unit 125 may store a consumer's search history, shopping cart, gender, age, fruit taste, preference, eco-friendly fruit, and allergy information, respectively.

Allergy information stored in the consumer database unit 125 may receive genetic information such as blood type information, disease name presence, family disease history, etc. received from the consumer terminal 10, and the genetic information collected from the consumer terminal 10 It is also possible to classify the consumer's gene type based on it. The consumer database unit 125 may further include a genetic information collection unit (not shown). The genetic information collection unit may output various inquiries for collecting genetic information to the consumer through the consumer terminal 10 . Accordingly, the customer's genetic information can be collected based on the responses inputted by the customer. Of course, allergy information may be directly input from the consumer terminal 10 .

When the fruit selection unit 140 selects a fruit according to the consumer's search history and shopping cart information stored in the consumer database unit 125, a mapping table in which shipment information of each fruit stored in the fruit characteristic database unit 122 is stored is referred to. can In addition, when the fruit selection unit 140 selects and excludes fruits based on the allergy information and genetic information stored in the consumer database unit 125, referring to the mapping table of the fruit characteristic database unit 122, the same allergy can be induced. You can select by excluding available fruits.

The purchase purpose database unit 126 stores the purchase purpose of the consumer. The purchase purpose of the consumer is the use of diet collected by the information collection unit 110, the use for the test taker, the use for the employment preparation, the use of the filial piety for the parents, the use of the housewarming, the use of a gift in return, the use of the visit, the use of the health care and the disease Therapeutic use and the like may be included. When the fruit selection unit 140 selects a fruit according to the purchase purpose of the consumer, the mapping table for the efficacy and characteristics of each fruit stored in the fruit efficacy database unit 121 and the fruit characteristics database unit 122 may be referred to.

Referring back to FIG. 1 , the AI curation-based fruit recommendation system according to the present invention further includes a big data analysis unit 130 . The big data analysis unit 130 may derive a consumer's transaction pattern and the like through big data analysis of each piece of information stored as a mapping table in the database unit 120 .

The big data analysis unit 130 analyzes in connection with fruit information and consumer information including the efficacy and characteristics of the fruit. The big data analysis unit 130 is collected by the information collection unit 110 and stored in each database unit 120 for fruit efficacy, characteristics, price, shipping time, consumer information, purchase purpose, and consumer transaction pattern (hereinafter, 'transaction patterns, etc.') can be derived. The transaction pattern, etc. derived by the big data analysis unit 130 may be used when predicting the purchase date of the consumer.

The big data analysis unit 130 may standardize a transaction pattern and the like and use it as an input of an unsupervised learning-based pattern calculation model. Standardization means, for example, quantifying the information collected from the producer terminal 20, such as the efficacy, characteristics, and price of fruit, and the information collected from the consumer terminal 10, such as consumer information, transaction patterns, and purchase purpose. do. A weight may be applied when quantifying information collected from the producer terminal 20 and the consumer terminal 10 . For example, a double weight is applied to information collected over the last 3 years among the information collected from the consumer terminal 10 , and a weight of 1 time is applied to information collected 4 to 7 years ago thereafter, and 7 years A weight of 0.5 times may be applied to previously collected information. However, the present invention is not limited thereto.

The big data analysis unit 130 may derive a transaction pattern or the like based on a clustering algorithm based on unsupervised learning.

Unsupervised learning refers to an algorithm that learns while analyzing or clustering data itself, rather than constructing learning data. Unsupervised learning because the fruit selection unit can learn the trade patterns of fruits by collecting data to be traded or traded in real time as well as the established past data when selecting and excluding fruits according to the trade patterns of fruits, etc. It may be more effective to derive a transaction pattern or the like based on a clustering algorithm based on it.

The big data analysis unit 130 may cluster and calculate purchase patterns and history patterns based on a clustering algorithm, and may detect new purchase patterns and history patterns based on the degree of separation between clusters of purchase patterns and history patterns. . For example, as a clustering algorithm for unsupervised learning, a logistic regression algorithm, a random forest algorithm, a support vector machine (SVM) algorithm, a decision-making algorithm, and a clustering algorithm may be used. In addition, the big data analysis unit 130 performs clustering algorithms such as Extra Tree Algorithm, XG Boost Algorithm and Deep Learning Algorithm, K-means Clustering Algorithm, SOM (Self-Organizing-Maps) Algorithm, EM & Canopy Algorithm in addition to the above-mentioned algorithms. Unsupervised learning can be performed through The Random Forest algorithm is an algorithm that consists of a forest of numerous decision trees and averages each prediction result into one result variable. It is a non-stochastic algorithm. Extra Tree Algorithm is similar to Random Forest but faster than Random Forest. XGBoost Algorithm is a boost algorithm that applies the result of XGBoost Tree to the next tree if the Tree of Random Forest is independent. The deep learning algorithm is an algorithm that learns by controlling the influence of variable patterns on the results with weights based on a multi-layered neural network. In addition, the K-means clustering algorithm is a traditional classification technique that iteratively subdivides the target group into K clusters based on the average value (similarity) of distance. It is a learning and clustering technique. In addition, the EM & Canopy algorithm refers to a method of clustering by updating parameter values through an iterative process starting with the maximum possibility with a given initial value.

The transaction pattern derived by the big data analysis unit 130 may include annual, monthly, weekly and hourly patterns. Fruits are usually shipped at a fixed time. The time of shipment of fruits can be largely divided into seasons. Therefore, deriving the transaction pattern as a monthly pattern may be effective in predicting the fruit transaction pattern. In addition, in the case of fruit, it can greatly depend on the temperature of the area where the fruit farm is located, and the average annual temperature is increasing every year due to current global warming. Therefore, deriving the transaction pattern as an annual pattern may be effective in predicting consumers' purchase of fruit by deriving the fruit transaction pattern. Deriving a transaction pattern by week may be effective when fruits are selected by the fruit selection unit 140 with reference to the producer's fruit shipment schedule and fruit inventory. It may be effective to predict a consumer's purchase time by deriving the transaction pattern as a pattern by time.

The fruit selection unit 140 selects fruits to be delivered to the consumer in a customized manner according to the information analyzed by the big data analysis unit 130 . The big data analysis unit 130 formalizes the information collected by the information collection unit 110 and stored in each database unit 120 . The information analyzed by the big data analysis unit 130 may be a transaction pattern or the like, as described above.

As an embodiment of the present invention, the fruit recommendation server 100 may receive a purchase amount from the consumer terminal 10 . The fruit recommendation server 100 may further include a price data storage unit (not shown) in which the purchase amount presented from the consumer terminal 10 is stored.

The fruit selection unit 140 may select a plurality of fruits and various kinds of fruits according to the purchase amount presented from the consumer terminal 10 and the consumer's transaction pattern.

When the fruit selection unit 140 selects multiple types of fruits, the big data analysis unit 130 may standardize the information stored in the database unit 120 and select a plurality of fruit species having a high numerical value. The number of types of fruit may be determined according to a preset value, but is not limited thereto, and may be determined as the number of an arbitrary number received from the consumer terminal 10 . That is, when the purchase price is presented from the consumer terminal 10 , the number of types of fruits may also be received. When the fruit selection unit 140 selects a variety of fruits, a greater weight may be given to the consumer database unit 125 in which the consumer's search history, fruit taste, preference, and allergy information are stored, respectively. For example, by comparing the mapping table in which the consumer's allergy information is stored and the mapping table in which the allergy information of each fruit stored in the fruit characteristic database unit 122 is stored, if there is a duplicate fruit, the fruit may be excluded from selection. . Also, for example, when the consumer's transaction pattern mainly deals with fruits having sour characteristics, the fruit may be selected by weighting the sour taste from the fruit taste mapping table stored in the fruit characteristic database unit 122 . Also, the fruit selection unit 140 may select a fruit with reference to the shipping time and stock information of the fruit.

When the fruit selection unit 140 selects a plurality of fruits, the fruit price database unit 123 is referred to, and the suggested purchase amount is weighted on the packaging unit of the fruit, the average weight per piece, and the selling price per unit weight information, and the number of fruits can be selected.

Fruit recommendation server 100 according to the present invention may further include an identification number providing unit. The identification number providing unit provides identification numbers to producers and consumers. The producer terminal 20 and the consumer terminal 10 can be registered as members in the fruit recommendation server 100 through a communication network, and a unique identification number is provided to customers who are registered as members at the time of initial registration, and thereafter The identification number provision procedure can be omitted.

The fruit recommendation server 100 according to the present invention may further include a natural language processing unit (not shown) for processing a signal received from the consumer terminal 10 in order to receive the purchase price suggestion in a natural language. The consumer terminal 10 may present the purchase amount by transmitting a text including the purchase amount to the fruit recommendation server 100 by a text message, MMS, or e-mail method. Alternatively, text can be entered through SNS applications such as KakaoTalk, Line, Facebook, Twitter and Instagram. Furthermore, the inputted text may be transmitted through a predetermined individual application. In addition, the purchase amount may be presented through a specific website. The fruit recommendation server 100 may further include a text storage unit (not shown) for storing the text received from the consumer terminal 10 . The natural language processing unit performs natural language processing on the text stored in the text storage unit.

In the present specification, text refers to all types of character information that can be recognized by a user, and refers to character information including all ranges of words, phrases, sentences, paragraphs, etc., including phoneme or syllable information .

Meanwhile, the fruit recommendation server 100 according to the present invention may further include a text processing unit (not shown) for recognizing sentences and words from natural language-processed text. The text processing unit may interpret the text in such a way that a meaning is given to each word by analyzing the relationship between the words in the recognized sentence.

As an additional explanation for natural language processing, natural language processing is a natural language (natural language) understanding or reconstruction of natural language (natural language) that is made into a form that can be understood by a computer by mechanically analyzing the linguistic phenomena uttered by humans. It refers to various technologies expressed in a language that humans can understand.

Such natural language processing can be done through morphological analysis and part-of-speech attachment. For morpheme analysis, for example, 'complex noun' can be divided into various methods such as 'complex + noun', 'complex + compound noun', 'compound + noun', etc., and the most suitable decomposition result can be selected from among them. . In order to select an appropriate result from among the variously divided analysis results, a dynamic programming method called table parsing is used. Specifically, if the probability of splitting into 'compound + noun' is greater than the probability of other analysis, it is an algorithm that selects a form with a higher probability.

Part-of-speech attachment refers to attaching the most appropriate form of speech among the results obtained through morpheme analysis. A module, usually called a tagger, performs this function. This is a module that selects one analysis result suitable for the context from among various analysis results output by the morpheme analyzer. During analysis, an appropriate analysis result is selected using information that is a hint for disambiguation located on the left and right of the context. Usually, the tagger is implemented using a large-scale part-of-speech attachment, and the Hidden Markov Model (HMM) is widely used.

The text processing unit can recognize the text received from the consumer terminal 10 through natural language processing, and keywords (hereinafter, 'keywords' etc') can be recognized. The keyword or the like recognized by the text processing unit may be compared with nouns stored in each column and row of the mapping table included in the database unit 120 to recognize whether the same noun exists. For example, when the text “Send 10,000 won worth of bananas for diet” is received, the text processing unit is the purchase amount '10,000 won' among the transmitted texts, and the purchase purpose mapping stored in the purchase purpose database unit 126 It is possible to recognize the 'diet' included in the table and the 'banana' included in the fruit characteristic mapping table stored in the fruit characteristic database unit 122 . Accordingly, the fruit selection unit 140 may select the number of bananas corresponding to the presented purchase amount to meet the 10,000 won limit. In addition, when the frequency of purchasing peaches in the consumer's purchase history is high, the fruit selection unit 140 may also select peaches. Peaches are 39kcal per unit weight (100g), low in calories, and 89% of the fruit is water. However, it goes without saying that the selection of peaches can be excluded when the information on the peach allergy in the consumer's allergy information is stored in the database unit 120 .

The fruit recommendation server 100 according to the present invention may further include a voice recognition unit (not shown) that receives a voice signal including the purchase amount. The voice recognition unit may convert the received voice signal into text and store it in the text storage unit. The text storage unit, as described above, after the natural language processing is performed by the natural language processing unit, it is possible to recognize the text by the text processing unit. When the fruit recommendation server 100 receives a voice signal, the consumer terminal 10 may further include a microphone unit (not shown) capable of inputting a voice signal.

In addition to the above-described algorithm, various methods for text processing or speech recognition may be used, of course, and the present invention is not limited to the above-described algorithm.

Meanwhile, the text processing unit may further include a mood state unit (not shown) for recognizing the mood state of the consumer through keyword recognition. The mood state unit can recognize mood states including depression, tension, expectation, anger, joy and sadness from the text. On the other hand, the information collection unit 110 may collect information about fruits that are helpful for depression, tension, expectation, anger, joy and sadness, and the collected information is stored in the fruit efficacy database unit 121 . Therefore, after recognizing the mood state of the consumer in the text transmitted from the consumer terminal by the mood state unit, information useful for the mood state is referred to in the fruit efficacy database unit 121 for the recognized mood state keyword. The referenced mood state information may be used when fruits are selected or excluded by the fruit selection unit 140 .

For example, when the text processing unit receives the text "I am depressed, do you have anything to eat" as the text transmitted from the consumer terminal 10, the text processing unit recognizes the word 'depressed'. When the text processing unit recognizes the word 'depressed', the fruit selection unit 140 may select a fruit that is helpful when depressed. Bananas are rich in tryptophan, which is involved in the production of the happy hormone serotonin. The magnesium and potassium in bananas can help relax the muscles and calm the town. Bananas are also rich in B vitamins, which are effective in relieving fatigue and relieving stress. Accordingly, when the text processing unit recognizes the word 'depressed' among the text received from the consumer terminal 10, the recommendation unit (not shown) included in the fruit recommendation server 100 according to the present invention can recommend bananas. .

The fruit recommendation server 100 according to the present invention may further include a recommendation unit for recommending a specific fruit. The recommendation unit may recommend a specific fruit when a text in which a purchase amount is not presented is received. However, the present invention is not limited thereto, and for example, the fruit selection unit 140 may select a fruit recommended by the recommendation unit by the price included in the transaction pattern analyzed by the big data analysis unit 130 .

Meanwhile, the fruit recommendation server 100 according to the present invention may further include a recommendation display unit (not shown) for transmitting information and images of fruits recommended by the recommendation unit to the consumer terminal 10 . The recommendation display unit may transmit information and images of a plurality of fruits to the consumer terminal 10 . The consumer terminal 10 may select a specific fruit from among a plurality of received fruits, and may present a purchase amount. Accordingly, the fruit selected to match the presented purchase amount may be selected by the fruit selection unit 140 .

3 is a flowchart showing a flow according to the AI curation-based fruit recommendation method according to the present invention.

Referring to FIG. 3 , the AI curation-based fruit recommendation method according to the present invention includes the steps of receiving a purchase price from the consumer terminal 10 ( S101 ), and delivering the purchase amount to the producer terminal 20 . (S103), the step of receiving information about circulating fruit from the producer terminal 20 (S105), the step of analyzing the information collected from the consumer terminal 10 to extract the consumer's transaction pattern (S107), and the step of extracting the consumer's and selecting the fruit to be delivered to the consumer in a customized manner according to the transaction pattern and the purchase amount (S109).

Step S101 is a step of receiving a purchase amount from the consumer terminal 10 . When receiving the purchase amount presented in step S101, as described above, the consumer terminal 10 may present the purchase amount by transmitting a text including the purchase amount through a text message, MMS, or e-mail method. Alternatively, text can be entered through SNS applications such as KakaoTalk, Line, Facebook, Twitter and Instagram. Furthermore, the inputted text may be transmitted through a predetermined individual application. In addition, the purchase amount may be presented through a specific website. On the other hand, after being transmitted in the form of a voice signal, the text may be extracted by analyzing the voice signal.

The inputted or extracted text may include the purchase amount suggested by the consumer, and at the same time include the name of the fruit desired by the consumer or the consumer's mood. As described above, the text may be analyzed by the natural language processing unit and the text processing unit to derive keywords.

Step S103 is a step of transferring the purchase amount to the producer terminal 20 . In step S103, the fruit recommendation server 100 according to the present invention compares the keyword derived in step S101 with the purchase amount presented from the consumer terminal 10 and the information stored in the database unit 120, and the information calculated is also produced by the producer terminal. It can be forwarded towards (20). Meanwhile, the fruit recommendation server 100 may transmit a plurality of pieces of information received from a plurality of consumer terminals 10 together to the producer terminal 20 .

Step S105 is a step of receiving information about the fruit that can be distributed from the producer terminal 20 . The producer terminal 20 transmits to the fruit recommendation server 100 the characteristics of the fruit containing information such as the shipping time and stock of the fruit.

Step S107 is a step of analyzing the information collected from the consumer terminal 10 to extract the consumer's transaction pattern and the like. In step S107, the big data analysis unit 130 analyzes the information about the fruit and the information about the consumer stored in the database unit 120 to extract the consumer's transaction pattern and the like.

Step S109 is a step of selecting the fruit to be delivered to the consumer in a customized manner according to the extracted consumer's transaction pattern and the suggested purchase amount. In step S109, the keyword derived in step S101 may be referred to. In addition, in step S109, the consumer's transaction pattern extracted in step S107 and the information stored in the database are compared, and a variety and a plurality of fruits are selected according to the received purchase amount. The selected fruit may be packaged and packaged, and the packaged fruit may be shipped to the consumer.

The fruit recommendation server 100 according to an embodiment of the present invention may communicate with the producer terminal 20 and the consumer terminal 10 through a communication network. The communication network refers to a connection structure capable of exchanging information between each node, such as terminals and servers. Examples of such a communication network include a 3rd Generation Partnership Project (3GPP) network, a Long Term Evolution (LTE) network, and 5G Network, WIMAX (World Interoperability for Microwave Access) network, Internet, LAN (Local Area Network), Wireless LAN (Wireless Local Area Network), WAN (Wide Area Network), PAN (Personal Area Network), wifi network, A Bluetooth network, a satellite broadcasting network, an analog broadcasting network, a Digital Multimedia Broadcasting (DMB) network, etc. are included, but are not limited thereto.

In addition, the producer terminal 20 and the consumer terminal 10 are terminals such as a desktop PC, a notebook PC, a tablet PC, and a smart phone equipped with input means such as a keyboard, a mouse, a touch pad, a touch screen, and a display screen. , but is not limited thereto, it is possible to access the fruit recommendation server 100 through a communication network, input search information and selection information, and process digital information capable of installing an application program that can display search result information Any configuration can be included.

The producer terminal 20 and the consumer terminal 10 are, for example, a desktop PC (desktoppersonal computer), a laptop PC (laptop personal computer), a netbook computer (netbook computer), a smartphone (Smartphone), a smartpad (SmartPad) , tablet personal computer (PC), mobile phone, video phone, PCS (Personal Communication System), GSM (Global System for Mobile communication), PDC (Personal Digital Cellular), PHS (Personal Handyphone System), PDA ( Personal Digital Assistant), IMT (International Mobile Telecommunication)-2000, CDMA (Code Division Multiple Access)-2000, W-CDMA (W-Code Division Multiple Access), Wibro (Wireless Broadband Internet) terminal, PMP (portable multimedia player) , all kinds of wireless communication devices such as wearable devices (e.g. smart glasses, head-mounted-device (HMD), etc.) or smart watches, and desktop computers, smart TVs, etc. It may include at least one of a wired communication device.

In this specification, a "part" includes a unit realized by hardware, a unit realized by software, and a unit realized using both. In addition, one unit may be implemented using two or more hardware, and two or more units may be implemented by one hardware.

In this specification, fruit recommendation server 100, communication unit, information collection unit 110, database unit 120, fruit efficacy database unit 121, fruit characteristics database unit 122, fruit price database unit 123, fruit Transaction database unit 124, consumer database unit 125, purchase purpose database unit 126, big data analysis unit 130, fruit selection unit 140, identification number providing unit, amount data storage unit, natural language processing unit, The text storage unit, the text processing unit, the voice recognition unit, the microphone unit, the mood state unit, the recommendation unit, and the recommendation display unit may be processors that execute consecutive execution processes stored in the memory. Alternatively, it may operate as software modules driven and controlled by a processor. Furthermore, the processor may be a hardware device.

For reference, the AI curation-based fruit recommendation method according to an embodiment of the present invention may be implemented in the form of a program command that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the present invention, or may be known and available to those skilled in the art of computer software. Examples of computer-readable media include hard disks, magnetic media such as floppy disks and magnetic tapes, optical recording media such as CD-ROMs and DVDs, magneto-optical media such as floppy disks, and ROM, RAM, A hardware device specifically configured to store and execute program instructions, such as flash memory, may be included. Examples of program instructions include not only machine code generated by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the present invention, and vice versa.

The protection scope of the present invention is not limited to the description and expression of the embodiments explicitly described above. In addition, it is added once again that the protection scope of the present invention cannot be limited due to obvious changes or substitutions in the technical field to which the present invention pertains.

10: consumer terminal
20: producer terminal
100: fruit recommendation server
110: information collection unit
120: database unit
121: fruit efficacy database part
122: fruit characteristic database unit
123: fruit price database part
124: fruit transaction database unit
125: consumer database unit
126: purchase purpose database unit
130: Big data analysis unit
140: fruit selection unit

Claims (5)

A fruit recommendation server capable of communicating fruit purchase information with a producer terminal and a consumer terminal, and recommending fruit suitable for a user's transaction pattern,
The fruit recommendation server,
an information collection unit for collecting information from the producer terminal and the consumer terminal;
a database unit in which the collected information is stored;
a big data analysis unit for extracting a consumer's transaction pattern through big data analysis of the information stored in the database unit; and
a fruit selection unit for selecting fruits to be delivered to the consumer in a customized manner based on the information analyzed by the big data analysis unit;
including,
The big data analysis unit,
Classifying the information stored in the database unit by year collected and applying a weight to the classified information to extract the consumer's transaction pattern,
The consumer's transaction pattern includes annual, monthly, weekly and hourly patterns,
The fruit recommendation server receives a purchase amount from the consumer terminal,
The fruit selection unit selects the fruit according to the received purchase amount and the consumer's transaction pattern,
The fruit recommendation server,
a text storage unit for receiving and storing text from the consumer terminal;
a natural language processing unit for natural language processing the text stored in the text storage unit;
a text processing unit for recognizing a noun from the text processed by the natural language processing unit;
further comprising,
The text processing unit,
From the natural language-processed text, keywords including the fruit name, the efficacy of the fruit, the characteristics of the fruit, the suggested purchase price or the fruit purchase purpose are recognized, and the keyword is compared with the information stored in the database unit, and the fruit selection unit AI curation-based fruit recommendation system, characterized in that it generates basic information when selecting a fruit. The method according to claim 1, The information collection unit,
Collecting efficacy, characteristics, and price information of each of various types of fruit from the producer terminal,
AI curation-based fruit recommendation system, characterized in that collecting transaction information and consumer information from the consumer terminal. The method according to claim 2, The database unit,
from the producer terminal,
a fruit efficacy database unit in which anticancer, antidepressant and antioxidant information of the fruit is stored, respectively;
a fruit characteristic database unit in which taste, color, nutrition, and shipping time information of the fruit are stored, respectively; and
and a fruit price database unit in which information on the fruit packaging unit of the fruit, the average weight per piece, the purchase price per unit weight and the selling price per unit weight information is stored, respectively;
from the consumer terminal,
a fruit transaction database unit in which the consumer's purchase frequency, purchase amount, and purchase satisfaction information are respectively stored;
a consumer database unit for storing the consumer's search history, gender, age, fruit taste, preference and allergy information, respectively; and
a purchase purpose database unit in which the purchase purpose of the consumer is stored;
AI curation-based fruit recommendation system comprising a. delete An AI curation-based fruit recommendation method comprising a fruit recommendation server capable of communicating fruit purchase information with a producer terminal and a consumer terminal and recommending a fruit suitable for a user's transaction pattern,
The method is
receiving, by the fruit recommendation server, a purchase amount from a consumer terminal;
transmitting, by the fruit recommendation server, the purchase amount to the producer terminal;
receiving, by the fruit recommendation server, information on available fruits from the producer terminal;
extracting, by the fruit recommendation server, a transaction pattern of a consumer by analyzing the information collected from the consumer terminal; and
selecting, by the fruit recommendation server, fruit to be delivered to the consumer in a customized manner according to the extracted transaction pattern of the consumer and the purchase amount;
including,
The step of the fruit recommendation server receiving the purchase amount from the consumer terminal,
The fruit recommendation server receives and stores the text from the consumer terminal, natural language processing the stored text, recognizing a noun from the natural language processing text, and fruit name, fruit efficacy, and fruit characteristics from the natural language processing text , recognizes a keyword including a suggested purchase amount or purpose of purchasing fruit, compares the keyword with information stored in a database unit, and generates information that is a basis for selecting a fruit,
The fruit recommendation server extracts the consumer's transaction pattern by analyzing the information collected from the consumer terminal,
The fruit recommendation server categorizes the collected information by collected year and applies a weight to the classified information to extract the consumer's transaction pattern, and the consumer's transaction pattern is a pattern for each year, month, week, and hour. AI curation-based fruit recommendation method, characterized in that it includes.

Images