AI Personal Fragrance Consultation And Fragrance Selection / Recommendation

By building an artificial intelligence system that uses sensors and machine learning to analyze user characteristics and generate personalized fragrance recommendations, the problem of matching individual fragrance preferences in existing technologies is solved, thus improving the accuracy of recommendations and user satisfaction.

CN119013682APending Publication Date: 2024-11-22COTY INC
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202380033733.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2022-11-28
Filing Date
2023-02-15
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

Existing technologies struggle to effectively identify and recommend products that match individual fragrance preferences, hindering marketing and personalized formulation.

Method used

By building an artificial intelligence system that uses sensors, cameras, microphones, and input devices to collect user data, and combining it with machine learning models to analyze user characteristics and a fragrance database, personalized fragrance recommendations are generated.

Benefits of technology

It enables precise fragrance recommendations based on factors such as user personality, mood, and environment, thereby improving user satisfaction and marketing effectiveness.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119013682A_ABST
    Figure CN119013682A_ABST
Patent Text Reader

Abstract

Data relating to a plurality of individuals is collected, and the data is used to identify whether any one of the data is a possible predictor of an individual's preference for one or more fragrances. These predictors are then used to provide a recommendation of one or more fragrances to one or more individuals.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] Priority declaration

[0002] This patent application claims the benefit of priority to U.S. application serial number 63 / 310,493 filed on February 15, 2022 and U.S. application serial number 63 / 385,082 filed on November 28, 2022, which applications are incorporated herein by reference in their entirety. Technical Field

[0003] The present disclosure relates generally to fragrances, and more particularly to recommendations for one or more fragrances. Background Art

[0004] How scent is perceived and associated is very subjective and varies widely between individuals. This makes it particularly difficult to recommend fragrances for products such as perfumes or scented products.

[0005] Further complicating the subjective nature of fragrance preference is that, in addition to each individual's own memories and experiences, it can also be influenced by their mental or emotional state. Scent has long been recognized to have a strong association with memory.

[0006] Given the wide range of variables and scents, a method is needed to identify the most likely predictors of fragrance preference and additionally the ability to generate recommendations based on the identified predictors.

[0007] The difficulty of matching fragrances to personal preferences makes it difficult to market products that use fragrances. The ability to match fragrances to user preferences can help create new or personalized formulations of fragrances, market scented products, and may also prove useful in some therapeutic applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 : According to an embodiment, a system for making fragrance recommendations is shown.

[0009] Figure 2 : According to an embodiment, a fragrance database is shown.

[0010] Figure 3 : According to an embodiment, a customer database is shown.

[0011] Figure 4 : According to an embodiment, an associative database is shown.

[0012] Figure 5 : According to an embodiment, a basic module is shown.

[0013] Figure 6 : According to an embodiment, a data collection module is shown.

[0014] Figure 7 : According to an embodiment, an analysis module is shown.

[0015] Figure 8 : According to an embodiment, a parameter selection module is shown.

[0016] Fig. 9 : According to an embodiment, a recommendation module is shown.

[0017] Fig.10 : According to an embodiment, a fragrance selection module is shown.

[0018] Fig.11 : According to an embodiment, a feedback module is shown.

[0019] Fig.12 : According to an embodiment, fragrance is shown as a beauty and style proposition.

[0020] Fig.13 : According to an embodiment, components for fragrance recommendation are shown.

[0021] Fig.14 : According to an embodiment, an example of a workflow for making a fragrance recommendation is shown.

[0022] Fig.15 : According to an embodiment, an example of a gifting workflow for making a fragrance recommendation is shown.

[0023] Fig.16 : According to an embodiment, an example of a personality connection between a conversation initiator and a refinement trigger is shown.

[0024] Fig.17 : According to an embodiment, examples of refining trigger factors are shown.

[0025] Fig.18 : According to an embodiment, an example of customer to product mapping is shown.

[0026] Fig.19 : According to an embodiment, an example of a scent to personality database is shown.

[0027] Fig. 20A , 20B and 20C: showing a classification of personality traits, according to an embodiment.

[0028] Fig.21 : According to an embodiment, an example of keyword grouping is shown.

[0029] Fig. 22 is a flow chart of an example of a method for making fragrance recommendations according to an embodiment.

[0030] Fig.23is a block diagram illustrating an example of a machine upon which one or more embodiments may be implemented. DETAILED DESCRIPTION

[0031] Embodiments of the present disclosure will be described more fully below with reference to the accompanying drawings, in which like numbers represent like elements throughout the several figures, and exemplary embodiments are shown in the accompanying drawings. However, the embodiments of the claims may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. The examples set forth herein are non-limiting examples and are merely examples of other possible examples.

[0032] Provided herein is an artificial intelligence tool that helps consumers choose one or more fragrances (aka perfumes) that are suitable for them. For example, AI diagnosis based on personality and sensory responses (olfactory series), style (fashion) and lifestyle (social; emotional; prominent features; preferences; occasions / events) stimuli, and optionally combined with facial attributes (including but not limited to eyebrow shape (e.g., straight, hard angle, soft angle, round, thick, average thickness, sparse, far away, tight gaps between, thin, normal thickness, short or normal length), eye shape (e.g., wide eye distance, narrow eye distance, average eye distance, large, small, average size, upturned, downturned, average angle, almond-shaped, round, narrow eyes, single eyelids, double eyelids, drooping eyelids, deep eye sockets, amber, brown, blue, gray, green, other colors or eye colors), nose shape (e.g., wide, narrow, average width, long, short or average length), cheekbones (e.g., flat, high, low or round), face shape (e.g., rectangular, square, heart-shaped, round, triangular, round ... triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, triangular, tri The analysis of lip shape (e.g., angular, diamond, or oval) and / or lip shape (e.g., lower lip thick, upper lip thick, thin, full, wide, narrow, round, arched, downturned, average, or lip color)) can be combined to make fragrance recommendations, which can be correlated with traits (including but not limited to extraversion (e.g., positive emotions, excitement seeking, activity level, self-confidence, greed, and warmth), agreeableness (e.g., trust, sincerity, altruism, cooperative compliance, modesty, and empathy), openness (e.g., imagination, aesthetic interests, emotional orientation, experimentality, intelligence, and tolerance of diversity), conscientiousness (e.g., self-confidence, organization, conscientiousness, achievement pursuit, self-discipline, and caution), and emotional stability (e.g., anxiety, irritability, dysregulation, self-consciousness, depression tendencies, and fragility)) and sub-traits (all of which contribute to the individual's fragrance mapping) and can then be used to further refine the consumer's (one or more) fragrance recommendations.

[0033] As can be seen in the Figures and Appendix, keywords from the consumer query are matched to categories that lead to scent personalities and then to fragrance recommendation(s) / mappings.

[0034] One aspect provides an online fragrance consultation that provides personalized fragrance recommendations for each person's true self, combining personal narratives, olfactory and sensory triggers, fashion and style expressions. In one aspect, the language from the consultation provides extraordinary fragrance language and conventions to help luxury fragrance consumers find what they are looking for, no matter how they express it and their culture, gender or dreams. In another aspect, the online consultation generates a (retailer) gift card or plan, providing the donor with the opportunity to provide an online fragrance consultation using a gift card.

[0035] This is a system for making fragrance recommendations. This system includes an electronic device 102 that is a computing device, which may include any of a mobile device, a phone, a tablet computer, a laptop computer, a desktop computer, a kiosk, etc. The electronic device 102 may include a general computing device or a specially constructed proprietary computing device, such as a kiosk or terminal that may be present in a retail location. The electronic device 102 may include at least one of a sensor 104, a camera 106, a microphone 108, or an input device 110. The electronic device 102 may include a communication interface 112 configured to communicate with the Internet or a cloud 114. The communication interface 112 may be connected to one or more electronic devices 102, such as a wearable mobile device including a smart watch or a wearable sensor 104. The communication interface 112 may also be connected to one or more cameras 106 or microphones 108. For the purposes of the present invention, any such sensor 104, camera 106, microphone 108, or input device 110 that can be connected to the electronic device 102 via the communication interface 112 will be referred to as a component of the electronic device 102, however, any such component can alternatively be one or more separate remote devices connected to the electronic device 102 via the communication interface. The sensor 104 is a detection or measurement device configured to collect data. The sensor 104 usually measures and quantifies analog input and converts it into digital data, although some sensors can collect and monitor digital data themselves. The sensor 104 can be any of a position sensor (accelerometer, global positioning system, etc.), a pressure sensor (pressure gauge, barometer, etc.), a temperature sensor (bolometer, thermocouple, thermometer, etc.), a force sensor (force transducer), a vibration sensor, a piezoelectric sensor, a fluid property sensor, a humidity sensor, a strain gauge, a photoelectric sensor, a flow switch, a level switch, and may also need to contact the item, substance, or material they are measuring, or may not need to contact. Similarly, some sensors can measure rotational motion and linear motion. The contactless sensor may additionally include a Hall effect sensor, a capacitive sensor, an eddy current sensor, an ultrasonic sensor, a laser sensor, or a proximity sensor. The sensor 104 may additionally include a consumable or catalytic chemical reaction, including an assay. Additional embodiments may include biometric monitoring sensors, such as a pulse oximeter, a galvanic skin response sensor, blood pressure, an electrocardiogram (EKG), and the like. The camera 106 is an imaging sensor or an array of imaging sensors, which typically measures reflected light and is then used to recreate and image on a display based on the measured value. Each measured value is used to fill a value into a pixel or sub-pixel. Multiple sub-pixels can create a complete pixel, and a pixel array creates an image. In some embodiments, multiple sensor measurements can be used to fill a single pixel or sub-pixel. Multiple sensor measurements can be acquired from one imaging sensor over a period of time, or simultaneously or also over a period of time from multiple imaging sensors.Multiple measurements may be averaged together, or multiple measurements may be subjected to a smoothing algorithm to determine a pixel or sub-pixel value. Each pixel or sub-pixel value may be determined independently, or may be determined via image processing of a portion or the entire image, the image consisting of multiple measurements in an array, to which one or more algorithms, such as smoothing, edge detection, etc., may be applied. Microphone 108 is an audio input device that detects sound waves and converts analog signals into digital data. Microphone 108 may be integrated into electronic device 102, a wearable device, or may be an independent microphone connected to an electronic device wirelessly or via a cable. Examples of microphones include condenser microphones, dynamic microphones, electret microphones, and the like. Microphone 108 may include a single audio pickup, or may include multiple audio pickups. The microphone may be integrated into electronic device 102 or connected via a cable or wirelessly via communication interface 112. Input device 110 is any device for capturing input from a user, such as an array of a keyboard, keypad, mouse, remote control, joystick, or any other switches, dials, etc. arranged to receive input from a user. The input device 110 may be further configured to capture gestures, for example, via a wearable device worn by a user or an image capture system for capturing and analyzing images and / or videos to detect gestures made by a user. The input device 110 may further include a touch screen interface, such as a capacitive, resistive, or pressure detection surface, which may or may not be covered on or below a screen capable of displaying content to a user. The input device 110 may also include a stylus. The communication interface 112 provides a connection between one or more electronic devices 102 or components. The communication interface 112 may have a physical interface to accept a cable connector such as an Ethernet cable, an optical cable, a USB cable, etc., or may provide a wireless connection. In order to provide a wireless connection, the communication interface 112 will include an antenna to send and / or receive data via electromagnetic waves. The wireless connection can be established using any communication protocol, such as Wi-Fi, Bluetooth, infrared (IR), cellular (3G, 4G, 5G, LTE, etc.), near field communication (NFC), radio frequency identification (RFID), global positioning system (GPS), etc. In some embodiments, the communication interface 112 can use light to establish a physical connection, such as using an optical cable or establishing wirelessly via one or more laser visible light communications. Cloud 114 is a network of distributed computing and data storage resources. Cloud 114 can be a public cloud, such as accessible via the Internet, or it can be a private cloud, which can be isolated and cannot be accessed via the Internet. Similarly, cloud 114 can be widely accessible, or access can be restricted through encryption, authentication, etc. In some embodiments, cloud 114 can be maintained by a third party, where resources are provided for one or more users and / or organizations.The fragrance database 116 stores data related to the names, descriptions, properties, ingredients, and chemical compositions of a variety of fragrances or compounds used to add fragrance and / or odor to products, such as perfumes, colognes, candles, air fresheners, shampoos, shower gels, deodorants, personal care products, detergents, etc. The data may additionally include manufacturers and information related to the manufacture of fragrances. The data may be populated by one or more manufacturers, suppliers, such as via a third-party database connected via the cloud 114. The customer database 118 stores data about one or more customers. The data may include characteristics that describe the user in general, such as always applicable, or conditionally describe the user's characteristics, such as sometimes only applicable under certain conditions. General characteristics may include the customer's personality, genetic information, allergies, customer-stated preferences, unless indicated as conditional, and the like. Conditional characteristics may include the customer's mood, occasion, location, or environment, and the like. The customer database 118 is populated by the data collection module 124 and the feedback module 134, and may additionally be populated by one or more proprietary or third-party databases and one or more application programming interfaces (APIs). In some embodiments, customer data may be stored in accordance with privacy and / or customer data retention policies and / or compliance with regulations such as the General Data Protection Regulation (GDPR). The association database 120 stores data related to the relationship between one or more customer characteristics and one or more fragrances. The association data indicates the possibility of one or more customer characteristics predicting the preference of one or more customers for one or more fragrances. The association database 120 is populated by the analysis module 126 and used by the parameter selection module 128 and the recommendation module 130. The base module 122 initiates the data collection module 124, which collects data from at least one data source including any one of a sensor, a camera, a microphone, an input device, a third-party database, a proprietary database, a survey, etc., and saves the collected data to the customer database 118. The base module 122 receives the collected data from the data collection module 124 and initiates the analysis module 126. The analysis module 126 uses the collected data and the data stored in the customer database 118 and the fragrance database 116 to identify one or more associations or customer characteristics and fragrances that are saved to the association database 120. The base module 122 receives the associations from the analysis module 126 and initiates the parameter selection module 128. The parameter selection module 128 evaluates each association to determine whether the association is a strong predictor of preference for one or more fragrances, and the parameter selection is saved to the association database 118. The base module 122 receives the selected parameters from the parameter selection module 128 and initiates the recommendation module 130, which creates and / or trains a recommendation model for generating one or more fragrance recommendations. The base module 122 receives one or more fragrance recommendations from the recommendation module 130 and initiates the fragrance selection module 132.The fragrance selection module selects one or more fragrances from the recommended fragrances identified by the recommendation module 130. The base module 122 receives (one or more) selected fragrances from the fragrance selection module 132 and initiates the feedback module 134. The feedback module 134 receives feedback from one or more customers and saves the feedback to the customer database 118. The base module 122 receives customer feedback from the feedback module 134 and ends the recommendation process. The data collection module 124 collects data from at least one source including any input device 110, and the input device is, for example, a camera, a sensor, a third party or a proprietary database, etc. The data collection module 124 initializes the input device 110 and receives data from the device, and then saves it to the customer database 118. The analysis module 126 receives data from the customer database 118 and the fragrance database 116. The analysis module 126 selects customer characteristic data from the customer database 118 and selects fragrance data from the fragrance database 116, and identifies the association of the pairing of customer characteristics and fragrances. In some embodiments, the association is quantifiable, while in other embodiments, the association is subjective or can be performed by a human screener. The association is saved to the association database 120, and the process is completed for all combinations of customer characteristics and fragrances. The parameter selection module 128 receives data from the fragrance database 116 and the association database 120, and for example, the ability of the association data to predict the preference of each fragrance is evaluated by comparing the quantified correlation coefficient with a threshold value and selecting the correlation as a parameter. Alternatively, the association with the highest correlation coefficient can be selected as a parameter. These parameters will be used to create a recommendation model, which will be used to select one or more fragrances to recommend to one or more customers. The recommendation module 130 receives the selected parameters determined by the parameter selection module 128 and the data from the customer database 118 to train the recommendation model. In an embodiment, the recommendation model can be a machine learning model. In an alternative embodiment, the recommendation model can include a lookup table, a decision tree, etc. The recommendation model is then used to generate recommendations based on a variety of available fragrances stored in the fragrance database 116, and the recommendation can include a preference score representing the probability that a customer reacts positively to the fragrance. The recommendation can include a collection or combination of fragrances, a category of fragrances, etc. The fragrance selection module 132 receives the multiple recommendations generated by the recommendation module 130, and additionally receives data from the fragrance database 116 and the customer database 118 for selecting one or more fragrances. In addition, when selecting one or more fragrances, customer characteristics, preferences or other weightings may be considered. The feedback module 134 receives the selected fragrance and sends a feedback request to the customer. The feedback request may be included in a sample of the recommended fragrance. Alternatively, the feedback request may include a digital survey request.The feedback module 134 also receives feedback from one or more customers, which feedback may include completing a survey, or may include behavioral prompts, such as purchasing a product containing one or more recommended fragrances, returning a product containing a fragrance, recommending a fragrance or product containing a fragrance to a social connection, or performing facial analysis of the customer's reaction when using or encountering a fragrance or a product containing a fragrance.

[0036] Now refer to Figure 2 Explain the functionality of the "fragrance database". Those skilled in the art will appreciate that for this process and method and other processes and methods disclosed herein, the functions performed in the process and method may be implemented in a different order. In addition, the steps and operations outlined are provided only as examples, and some of these steps and operations may be optional, combined into fewer steps and operations, or expanded into additional steps and operations without departing from the essence of the disclosed embodiments.

[0037] This figure shows a fragrance database 116. The fragrance database 116 stores data related to one or more fragrances. The data at least includes a fragrance name and a unique identifier or ID. The fragrance database 116 may additionally include a variety of fragrances or descriptions, characteristics, ingredients, and chemical compositions of compounds for adding fragrance and / or odor to products, such as perfumes, colognes, candles, air fresheners, shampoos, shower gels, deodorants, personal care products, detergents, etc. The data may additionally include information about manufacturers, brands, and supplier information and information related to the manufacture of fragrances, including manufacturing process steps. The data in the fragrance database 116 may be filled by one or more fragrance manufacturers, suppliers, etc., and used by an analysis module 126, a parameter selection module 128, a recommendation module 130, and a fragrance selection module 132.

[0038] Now refer to Figure 3 Explain the functionality of the "Customer Database". Those skilled in the art will appreciate that for this and other processes and methods disclosed herein, the functions performed in the processes and methods may be implemented in different orders. In addition, the steps and operations outlined are provided only as examples, and some of these steps and operations may be optional, combined into fewer steps and operations, or expanded into additional steps and operations without departing from the essence of the disclosed embodiments.

[0039] This figure shows a customer database 118. The customer database 118 stores data about one or more customers. The data may include identifiable information, such as name, phone number, email address, home address, etc., and may include characteristics that describe the user. In addition to regulations such as the General Data Protection Regulation (GDPR), customer data, especially personally identifiable information, should also be stored in compliance with privacy and customer data retention policies. Such policies and regulations may require the consent of the customer before the data is captured, stored, or sent to the server for analysis. In some embodiments, data can be collected and stored locally on the customer's private device without obtaining permission, but such permission may still be required before analyzing the obtained data. In the absence of receiving the customer's consent, the obtained data (including but not limited to the analysis results) analyzed using the private device cannot be sent outside the device. In some embodiments, consent can be implicit, so that a message is presented to the user, that is, continuing to use the device and / or program is deemed to agree to use their personal information, while in other embodiments, consent is explicit, requiring the customer to make a choice to confirm or refuse consent. Examples of characteristics may include gender, hobbies, personality traits, fragrance preferences, purchased items, feedback related to purchased products, etc. The data may additionally include information such as genetic information, allergies, non-fragrance preferences, emotions, personality, reactions to stimuli, occasions of interest, places, environments, etc. Some customer characteristics, such as personality, emotions, reactions, etc., may be determined based on analysis of other customer characteristics such as facial features, intonation, etc. These features may be used directly as customer characteristics and / or may be used to derive intermediate customer characteristics, such as personality, emotions, reactions, etc. The customer database 118 is populated by the data collection module 124 and the feedback module 134, and may additionally be populated by one or more proprietary or third-party databases and / or application programming interfaces (APIs). Examples of connected third-party data sources may include social media networks, supplier websites, search engines, web tracking, and advertising services, etc. The customer database 118 is used by the analysis module 126, the recommendation module 128, and the fragrance selection module 132. In some embodiments, the feedback module 134 may additionally use the customer database 118 to determine where to send feedback requests, such as customer satisfaction surveys and additional promotions.

[0040] Now refer to Figure 4 Explain the functionality of the "associative database". Those skilled in the art will appreciate that for this and other processes and methods disclosed herein, the functions performed in the processes and methods may be implemented in different orders. In addition, the steps and operations outlined are provided only as examples, and some of these steps and operations may be optional, combined into fewer steps and operations, or expanded into additional steps and operations without departing from the essence of the disclosed embodiments.

[0041] This figure shows an association database 120. The association database 120 stores data related to the relationship between one or more customer characteristics and one or more fragrances. The association data indicates the degree to which one or more customer characteristics are associated with one or more fragrances, such as indicated by a correlation coefficient. Other descriptors of the association can be descriptive rather than quantitative, such as high, medium, or low. Other examples can include unlikely, slightly unlikely, irrelevant, slightly likely, or likely to have a predictive association, so that the customer characteristic can indicate or be used to predict a preference for one or more fragrances. The association database 120 is generated by the analysis module 126 and used by the parameter selection module 128.

[0042] Now refer to Figure 5 The functions of the "base modules" are explained. Those skilled in the art will appreciate that for this process and method and other processes and methods disclosed herein, the functions performed in the process and method may be implemented in different orders. In addition, the steps and operations outlined are provided only as examples, and some of these steps and operations may be optional, combined into fewer steps and operations, or expanded into additional steps and operations without departing from the essence of the disclosed embodiments.

[0043] This figure shows the base module 122. The process begins by initiating the data collection module 124 at step 502. The data collection module 124 initializes the sensor 104, camera 106, microphone 108 and other input devices 110, and polls the initialized devices to find data. The data collection module 124 additionally accesses third-party data such as from social media, and prompts the user to provide direct input via at least the input device 110. The data is saved to the customer database 118. At step 504, the collected data is received from the data collection module 124. Data can be collected from any one of one or more sensors 104, cameras 106, microphones 108, input devices 110, or third-party sources, such as private or proprietary databases owned and maintained by social media companies. Initiate the analysis module 126 at step 506. The analysis module 126 queries the customer database 118 and the fragrance database 116 and selects the first fragrance and the first customer characteristics. Calculate the correlation between the first fragrance and the first customer characteristic and / or identify the association between the first fragrance and the first customer characteristic, and save the calculated correlation and / or the identified association to the association database 120. Further determine whether there are more customers, and if there are more customer characteristics available for analysis, select the second customer characteristic, otherwise determine whether there are more fragrances available. If there are more fragrances available, select the second fragrance to be analyzed. At step 508, when all customer characteristics and fragrance combinations have been analyzed, receive the analyzed data including the correlation and / or the identified association from the analysis module 126. The resulting data can be quantitative, such as a correlation coefficient indicating the possibility that one or more customer characteristics are predictors of the preference for one or more fragrances, or it can be a more general association between customer characteristics and fragrances. At step 510, initiate the parameter selection module 128. The parameter selection module 128 queries the fragrance database 116 and the association database 120, and selects the first fragrance and the first association, for example, by comparing the correlation coefficient with a threshold. Alternatively, the association between customer characteristics and fragrances can be identified by a method that does not rely on statistical principles including subjective evaluation. The strength of association of the association can be evaluated, and the context of the association can be additionally considered so that the combination of multiple associations can increase the predictive relevance of the association. If the correlation coefficient is above a threshold, or the association is found to be a sufficient predictor of fragrance preference, the correlation or association is selected as a parameter. If there are more associations, the second association is selected, otherwise if there are more fragrances, the second fragrance is selected. At step 512, the selected parameters are received from the parameter selection module 128. The selected parameters include correlations and / or associations that represent the pairing of customer characteristics and fragrances, which have a high probability of accurately predicting the customer's preference for one or more fragrances. At step 514, the recommendation module 130 is initiated.Recommendation module 130 queries customer database 118 and fragrance database 116 to find data relevant to selected parameters, and uses data training recommendation model including relevant customer characteristics and fragrance.In certain embodiments, recommendation model can be a lookup table or a decision tree.Recommendation model is also used to generate one or more fragrance recommendations.At step 516, receive one or more fragrance recommendations from recommendation module 130.Described recommendation can include fragrance, fragrance series or fragrance group of one or more shared common characteristics.At step 518, initiate fragrance selection module 132.Fragrance selection module 132 queries fragrance database 116 and customer database 118 to obtain recommended fragrance data and additional customer data, such as customer input and feedback, and select one or more fragrances from recommended fragrance.At step 520, receive one or more selected fragrances from fragrance selection module 132.At step 522, initiate feedback module 134.Feedback module 134 receives one or more selected fragrances, and optionally requests feedback from one or more customers. Customer feedback is received through direct or indirect methods, such as by providing surveys or monitoring sales activities, and the customer feedback is saved to the customer database 118. At step 524, customer feedback is received from the feedback module 134. The customer feedback includes data that explicitly or implicitly indicates one or more preferences for one or more fragrances. At step 526, the fragrance recommendation process ends.

[0044] Now refer to Figure 6 The functionality of the "data collection module" is explained. Those skilled in the art will appreciate that for this process and method and other processes and methods disclosed herein, the functions performed in the process and method may be implemented in a different order. In addition, the steps and operations outlined are provided only as examples, and some of these steps and operations may be optional, combined into fewer steps and operations, or expanded into additional steps and operations without departing from the essence of the disclosed embodiments.

[0045] This figure shows the data collection module 124. The process begins at step 602 from the base module 122 to receive a prompt to start data collection. The prompt can be an automatic or manual action to start collecting data from one or more customers. Data collection can be prompted by customer actions, such as registering on the website, purchasing products containing fragrances, etc. At step 604, one or more sensors 104 are initialized for data collection. Initialization includes powering on the sensor 104 device, and can additionally include handshaking, where a signal is sent to the sensor 104 and a response is received to confirm that the sensor 104 is powered on and ready to collect data. The sensor can additionally refer to a camera 106, a microphone 108, or any other input device 110. In an embodiment, a connection is established with a sensor 104 in a mobile phone, and the mobile phone includes an accelerometer, (one or more) microphones, (one or more) cameras, etc. of the phone. At step 606, one or more sensors 104, cameras 106, microphones 108, input devices 110, etc. are polled to find data input. In an embodiment, an accelerometer can collect location and movement data from a customer. In another embodiment, microphone 108 can collect audio data containing the voice of the customer, which can be analyzed later to determine the emotion of the customer. Similarly, camera 106 data can be collected. Other embodiments can collect data from sensor 104, which is configured to detect and analyze the composition of volatile compounds to determine the fragrance present when collecting data about the customer. Similarly, position data can be used to determine the position of the customer in the store, and determine which fragrance products or fragrance samples the customer may be near when collecting data that can be used to measure the reaction of the customer. In an embodiment, at least one camera 106 directed toward the customer captures at least one image of the customer when presenting at least one fragrance to the customer, so that the at least one image includes a facial expression that reacts to at least one fragrance. In other embodiments, there is no fragrance, but the customer can be prompted to follow the instruction to make an expression, which is captured by at least one camera 106. Such image data can be used to identify reactions, emotions or personality later, and the reactions, emotions or personality can be evaluated as potential predictors of fragrance preference. In another embodiment, the camera 106 captures the image data of the customer in order to perform facial feature analysis, such as shape, features and spacing, etc., which can be related to a set of personality traits and further used as a potential predictor of fragrance preference. Similarly, one or more microphones 108 can be used to collect audio samples of the customer's voice, the audio can be processed on the audio samples, and the tone of the customer's voice is analyzed to indicate a reaction, emotion or personality, which can be evaluated as a potential predictor of fragrance preference. The customer can also submit data via one or more input devices 110, for example, by following prompts and answering questions on a touch screen or entering data via a keyboard and / or mouse.At step 608, sensor 104 data is received from the sensor 104. The sensor 104 data may be temporarily stored in the sensor 104 device in batches or groups and sent to the electronic device 102, or may be continuously streamed. The method of data transmission may depend on the network connectivity at the time of data collection, so that when the network connection with the electronic device 102, the Internet or the cloud 114 is poor or no network connection, the data is stored on the local sensor 104 device and then sent when the network connection is available. Likewise, when the network connection is available, the data may be streamed in real time. In an embodiment, the electronic device 102 has a reliable connection with at least one sensor 104 such as an accelerometer, and receives data in real time. In an embodiment, image data is received in real time from at least one camera 106 directed toward the customer's face, thereby capturing facial expressions. The microphone 108 may be used to collect the customer's voice in addition. At step 610, third-party data from one or more third-party data sources is accessed. The third-party data may include a connected device, the connected device including a plurality of sensors 104, a camera 106, a microphone 108, and an input device 110. The third-party data source may additionally include a database owned and managed by a third party, such as a database managed by a social media provider. In an embodiment, the data collection module 124 accesses social media data from Facebook and Twitter, and determines friends, followers, and people followed by the customer, and additionally collects public posts that mention or tag the customer, as well as the customer's friends, followers, etc., published. Social media data may additionally include products and services that the customer has liked or otherwise expressed interest in. Third-party data may also include sales and return data of fragrances and clothing, cosmetics, personal care products, etc. from one or more suppliers. Third-party data may further involve location, geography, climate, weather, and calendar data, including data from public and private calendars. In an embodiment, a female customer connected to a retailer's database and confirming that a customer with customer ID 27046 purchased Her Eau de Parfum. At step 612, the user or customer is prompted to enter. The input prompt may include a text message, an instruction on a kiosk, a mobile device screen, a website, etc., and user input may be provided by an input device 110. User input may be provided by one or more sensors 104, a camera 106, a microphone 108, etc. in addition or alternatively. For example, the prompt may ask the customer to read text aloud or verbally indicate their response to be received via a microphone. In alternative embodiments, the user may be provided with an image or fragrance and their response may be captured via one or more cameras 106. Other embodiments may include completing the survey via a keyboard input device 110.The user input prompt may additionally include real-time video interaction between the customer and a video recording, or with a real person or dynamically generated content, or real-time video interaction using a pre-recorded video feed. At step 614, the data collected by the data collection module 124 is saved to the customer database 118. At step 616, return to the base module 122.

[0046] Now refer to Figure 7 Explain the functionality of the "Analysis Module". Those skilled in the art will appreciate that for this and other processes and methods disclosed herein, the functions performed in the processes and methods may be implemented in different orders. In addition, the steps and operations outlined are provided only as examples, and some of these steps and operations may be optional, combined into fewer steps and operations, or expanded into additional steps and operations without departing from the essence of the disclosed embodiments.

[0047] This figure shows the analysis module 126. The process begins with receiving a prompt from the basic module 122 at step 702. The prompt may contain customer data collected from the data collection module 124. The prompt may additionally include instructions for performing analysis on all data, only newly collected data, or only data related to a specific customer. At step 704, the customer database 118 is queried to obtain customer characteristic data from at least one customer. The data may additionally include customer feedback data, such as the customer's fragrance preferences, to include the fragrances they like, the fragrances they don't like, and may additionally include quantifiable scores, such as the customer describing the fragrance with ID 11, which has a score of 7 out of 10. Examples of customer characteristic data may include images of the customer's face, including when the customer reacts to a specific fragrance. The data may include voice and / or video recordings of the customer responding to the prompt. Customer characteristics may additionally include hobbies, personality traits, their gender, location, places visited, purchase history, etc. Any data that can be used to describe a customer, their preferences, behavior, etc. may be used as customer characteristics. Customer data may be aggregated data, or may be for a single customer. At step 706, the fragrance database 116 is queried to obtain one or more fragrances. The fragrance data may include specific fragrances, but may also include their characteristics, ingredients, etc., so that the analysis of the customer characteristics paired with the fragrances may be performed alternatively between the fragrance characteristics and / or ingredients. This allows for a broader comparison and recommendation, rather than only an explicit comparison. In some embodiments, both explicit data and general data are used to allow identification of a series of related or similar fragrances, from which recommendations can be selected for the customer. At step 708, fragrances, fragrance characteristics, or fragrance ingredients are selected from the data retrieved from the fragrance database 116. In an embodiment, a fragrance with ID 11 is selected, i.e. Her Eau de Parfum. At step 710, customer characteristics are selected from the customer data retrieved from the customer database 118. In an embodiment, customer characteristics are images of customers taken when customers sample fragrances with ID 11. Customers may have sampled fragrances different from the selected fragrances instead. Alternatively, customer characteristics may be unrelated to fragrances, such as the user's personality, mood, favorite food, favorite activity, etc. In another embodiment, an adventurous personality is selected as a customer characteristic. In some embodiments, customer characteristics may include customer categories or groups, which may be characterized by one or more customer characteristics. For example, people with an adventurous personality, or a preference for fruity fragrances, or people who like specific activities, such as hiking. Other groupings may involve demographics, habits, music preferences, clothing, hairstyle or makeup style preferences, preferred reading types, etc. In some embodiments, customer characteristics may include emotions or the emotional state of the customer.The emotional state of the customer can be identified based on facial expressions, postures, intonations, message contexts, or biometric information (e.g., heart rate, blood pressure, galvanic skin response, EKG, etc.). The emotional state can be identified using a method similar to personality by one or more algorithms including machine learning. At step 712, the association between the selected customer characteristics and the selected fragrance is identified. In an embodiment, the association is determined by calculating a correlation coefficient, which represents the probability that the selected customer characteristics are predictors of the customer's preference for the selected fragrance. In some embodiments, such as embodiments utilizing machine learning or another automated algorithm, the correlation coefficient can be a quantifiable statistical relationship, such as a Pearson correlation coefficient, which determines the dependence of a variable (in this case, the selected customer characteristics) on a second variable (in this case, the selected fragrance) by comparing data from a large number of customer samples. In alternative embodiments, such as embodiments using a decision tree or a lookup table, a more generalized association between the selected fragrance and the selected customer characteristics can be identified, such as a person with a specific customer characteristic is more likely or less likely to buy a specific fragrance. In some embodiments, the association can be determined based on interviews collected manually or digitally to determine customer preferences and create association definitions based on subjective or other analysis of the collected data. In an embodiment, the association may include gender (e.g., the customer is a female) and the associated fragrance Her Eau de Parfum. In another embodiment, a customer characteristic preference for a fruity fragrance is identified and associated with the fragrance My Burberry Eau de Toilette. In some embodiments, multiple customer characteristics may be considered together, such as a woman with a seductive personality that may be associated with the fragrance Her Eau de Parfum. The association represents a potential predictive relationship between one or more customer characteristics and one or more fragrances. At step 714, the association with the association database 120 is saved. At step 716, it is determined whether there are more customer characteristics to be analyzed. In an embodiment, there are additional customer characteristics to be analyzed, such as a customer's preference for kayaking. Since there are more customer characteristics, return to step 710 and select another customer characteristic. In an alternative embodiment, there are no additional customer characteristics to be analyzed. At step 718, it is determined whether there are more fragrances to be analyzed. In an embodiment, there are additional fragrances, fragrance characteristics, ingredients, etc. to be analyzed, such as the fragrance Brit For Her Eau de Parfum with ID 173. Since there are more fragrances to be analyzed, the process returns to step 708 and another fragrance is selected. In an alternative embodiment, there are no additional fragrances to be analyzed. At step 720, the process returns to the base module 122.

[0048] Now refer to Figure 8The function of the "parameter selection module" is explained. Those skilled in the art will understand that for this process and method disclosed herein and other processes and methods, the functions performed in the process and method may be implemented in different orders. In addition, the steps and operations outlined are provided only as examples, and some of these steps and operations may be optional, combined into fewer steps and operations, or expanded into additional steps and operations without departing from the essence of the disclosed embodiments.

[0049] This figure shows parameter selection module 128. The process begins at step 802 where prompts are received from base module 122. The prompts may include associated data created by analysis module 126. The prompts may additionally include instructions related to the type of recommendation engine to be created, for example, whether it will be used for a related fragrance group, or alternatively when a certain type of customer characteristic (e.g., interested in outdoor activities) is identified. At step 804, query fragrance database 116 to obtain one or more fragrances, fragrance properties, ingredients, etc. At step 806, query association database 120 to obtain at least one correlation or association between one or more fragrances, fragrance properties, ingredients, etc. and one or more customer characteristics. Association may be a correlation coefficient or any other number that quantifies the relationship between fragrance and customer characteristics or the probability that customer characteristics are predictors of customer preferences for fragrances. Association may additionally be a non-digital representation of the relationship between customer characteristics and fragrances, fragrance properties, ingredients, etc. At step 808, select fragrances, fragrance properties, or fragrance ingredients from the data retrieved from fragrance database 116. In an embodiment, a fragrance with ID11, Her Eau deParfum, is selected. At step 810, a correlation or association is selected from the data retrieved from the association database 120. The association includes at least a customer characteristic and an identified relationship with the selected fragrance. The identified relationship may include an indication of the strength of the relationship, which may be quantitative, such as in the case of a correlation coefficient, an ordinal number (e.g., a ranking), or subjective, including descriptors such as low, medium, or high. At step 581, it is determined whether the association is greater than a threshold. In an embodiment, it is determined whether the correlation coefficient is above a threshold. The threshold may be selected by a system administrator and then used to determine whether the customer characteristic represented by the correlation coefficient should be selected as a parameter. If the correlation coefficient is above a threshold, the association may be selected as a parameter, otherwise it will not be selected as a parameter. In some embodiments, a selected number of associations, such as up to 100, will be selected. In other embodiments, a maximum percentage, such as up to 10% of associations by the correlation coefficient, will be selected as a parameter. In alternative embodiments, a correlation coefficient may not be used, but a more generalized association may be used. For example, associations may be selected such as customers who like kayaking tend to prefer fragrance with ID 11, Her Eau de Parfum, while the same group tends not to prefer fragrance with ID 174, Brit For Her Eau de Toilette. These associations may be identified through direct observation of individual customers, such as by receiving survey responses indicating that they do not like such fragrances and other such feedback, or may be aggregated based on customer input and feedback from a large collection of collected data.In the case of aggregated data, if more than half of the relevant customers, such as customers identified as liking kayaking, like the fragrance with ID 11, an association can be identified. If less than half of the customers in the same group indicate that they prefer the fragrance, no association will be identified. The threshold can be moved, and in some cases, can be subjectively evaluated and assigned by a human screener. Other algorithms can be used to determine such associations. When associations are identified, they can be selected as parameters, otherwise they can be ignored. In some embodiments, the association can include multiple customer characteristics and / or fragrances. At step 814, if the correlation coefficient of the customer characteristic is higher than the threshold, or if there is a sufficiently important association that will be selected by manual selection or other selection algorithms or criteria, the customer characteristic represented by the correlation or association is selected as a parameter. For example, "liking kayaking" is selected as a parameter, which can be used to identify customers who prefer fragrances with ID 11. In some embodiments, a single parameter can be used to make recommendations. In other embodiments, multiple parameters can be used, such as in a decision tree or a lookup table. In another embodiment, a parameter for training a machine learning algorithm or modifying the weighting of a weighted value table or process to generate fragrance recommendations can be selected. In some embodiments, the parameters may include personality traits, emotions, reactions, etc., which may indicate a preference for the selected fragrance. Selecting the parameters and fragrances may include adding personality traits, emotions, reactions, etc. to a lookup table associated with the selected fragrance, or alternatively, the value weighting may be adjusted to determine whether the fragrance is recommended based on the parameter. At step 816, it is determined whether there are more associations to be analyzed. In an embodiment, there are additional associations to be analyzed, so return to step 810 and select another correlation or association. In an alternative embodiment, there are no additional correlations or associations to be analyzed. At step 818, it is determined whether there are more fragrances to be analyzed. In an embodiment, there are additional fragrances, fragrance properties, ingredients, etc. to be analyzed, such as the fragrance Brit For Her Eau de Parfum with ID 173. Since there are more fragrances to be analyzed, it is returned to step 808 and another fragrance is selected. In an alternative embodiment, there are no additional fragrances to be analyzed. At step 820, the selected parameters are returned to the base module 122.

[0050] Now refer to Fig. 9 Explain the functionality of the "recommendation module". Those skilled in the art will appreciate that for this process and method and other processes and methods disclosed herein, the functions performed in the process and method may be implemented in different orders. In addition, the steps and operations outlined are provided only as examples, and some of these steps and operations may be optional, combined into fewer steps and operations, or expanded into additional steps and operations without departing from the essence of the disclosed embodiments.

[0051] This figure shows the recommendation module 130. The process begins by receiving the selected parameters from the base module 122 at step 902. The selected parameters have met the quantitative threshold or have been selected due to the association with at least one fragrance preference identified. At step 904, the customer database 118 is queried to obtain the customer characteristics corresponding to the selected parameters. For example, the selected parameters can include the association between the customer's preference for kayaking and their preference for fragrances with ID 11, and the corresponding customer characteristics are the customer's preference for kayaking. In addition, data related to the interested customer for whom the recommendation is to be generated is retrieved. The data related to the interested customer can affect the training of the recommendation model by using only the parameters corresponding to the available data describing the customer. For example, if the customer's hobby data is available, the parameters related to the hobby preference will be included in the training data, and if the social media data is not available for the interested customer, the social media related parameters will be excluded from the training data. At step 906, the fragrance database 116 is queried to obtain the available fragrances. The fragrance data can include specific fragrances, but can also include their characteristics, ingredients, etc. Any such data describing fragrances can be used to generate a recommendation model. For example, fragrances can eventually be recommended as fragrance series, rather than specific fragrances. At step 908, a recommendation model is trained based on customer characteristics corresponding to the selected parameters. In machine learning applications, customer characteristics will correspond to the source data of features or mappings used to train machine learning models. Training machine learning models usually uses regression by applying adjustments or corrections after each continuous training test. Several evolutions can be completed, wherein a reserved selection of training data is retained as test data, so as to evaluate the accuracy of the trained model. Evolution can continue until the prediction accuracy of the model is higher than a threshold, such as 95%. The retained training data can be changed, so as to be randomly reselected between each evolution. The recommendation model can be alternatively based on a lookup table or a decision tree. In the lookup table, one or more customer characteristics based on the selected parameters can be used to directly map to one or more fragrances. In some cases, multiple lookup tables can be obtained using different combinations of selected parameters. For example, some tables can use personality traits, while other tables can use hobbies or interests. Other tables can include occasions, while other tables can include a combination of interests and occasions. These tables can be formed by mapping fragrances to customer characteristics using selected parameters. Decision trees can be created similarly to lookup tables, except that a series of branching decisions can be used instead instead of finding matches. For example, one decision tree can include the user's interests, while another decision tree can be occasions, and another decision can be the user's stated fragrance preferences. Similar to a lookup table, decision tree branching points can be determined based on a mapping of fragrances to customer characteristics. Selected parameters can be used to modify weights in a weighted value approach, such that a lookup table or algorithm is used where a matching customer characteristic can increase or decrease the likelihood of preference for one or more fragrances.At step 910, one or more fragrance recommendations are generated based on the trained recommendation model and the data collected for the client. The generated recommendation may include one or more specific fragrances, or may alternatively include one or more fragrance series so that the fragrances in the series have similar qualities. Similarly, the recommendation may be based on a specific ingredient or common property in the fragrance. In some embodiments, the recommendation is based on personality type. In other embodiments, the recommendation is based on the claimed preference of the client to a specific olfactory component or ingredient (e.g., "I prefer a floral scent like peony"), including the recommendation based on the client indicating that they currently use or prefer a specific fragrance. In another embodiment, the recommendation is based on the fragrance of new ingredients with similar olfactory attributes and characteristics. In another embodiment, the recommendation is based on ingredients similar to the ingredients in the current or preferred fragrance of the client. In other embodiments, the recommendation may use a weighted method (alone or in combination) so that the olfactory attributes or ingredient composition of the fragrance that the client has indicated that he prefers or currently uses are given more or less influence in the rating with preferences and characteristics (e.g., personality type, mood, occasion, etc.). Fragrance recommendation can include binary recommendation and non-recommendation decision, or can provide score in addition, for example, each recommended fragrance includes the possibility of customer preference (total score 10), and recommended fragrance can be any fragrance with score higher than threshold value (for example, 8 / 10). These scores can also be used to select one or more fragrances from recommended fragrances. In an embodiment, it is determined that the customer is a woman so that they are more likely to prefer Her Eau de Parfum. In another embodiment, it is identified that the customer characteristic preference of fruity fragrance shows that the preference for My Burberry Eau de Toilette increases. In some embodiments, multiple customer characteristics can be considered together, for example, women with seductive personality prefer Her Eau de Parfum. At step 912, one or more fragrance recommendations are sent to basic module 122.

[0052] Now refer to Fig.10 The functions of the "fragrance selection module" are explained. Those skilled in the art will understand that for this process and method and other processes and methods disclosed herein, the functions performed in the process and method may be implemented in a different order. In addition, the steps and operations outlined are provided only as examples, and some of these steps and operations may be optional, combined into fewer steps and operations, or expanded into additional steps and operations without departing from the essence of the disclosed embodiments.

[0053] This figure shows the fragrance selection module 132. The process begins at step 1002 where one or more fragrance recommendations are received from the base module 122. One or more fragrance recommendations may include specific fragrances or fragrance groups based on commonality. At step 1004, query the fragrance database 116 to obtain one or more fragrances corresponding to the fragrance recommendations received from the base module 122. For example, if a fragrance group is recommended, each unique fragrance in the fragrance group is identified as eligible to be selected. At step 1006, query the customer database 118 to obtain specific fragrance preferences or information about current recommendations. For example, if a customer wishes to purchase new fragrances for an upcoming party and / or prefers mild fragrances that will not conflict with the fragrance or flavor of dinner. At step 1008, fragrances are selected from fragrance recommendations based on the customer's preferences. In some embodiments, multiple fragrances may be selected. Fragrances may be selected in an automatic manner using one or more algorithms, or may be selected manually by a contractor or a customer instead. In some embodiments, multiple fragrances may be selected with the purpose of mixing them into new fragrances. In alternative embodiments, multiple fragrances may be selected with the goal of sending multiple fragrances to the customer rather than a single fragrance. In some embodiments, fragrances may be selected using customer data and a weighted value approach, such as a lookup table or algorithm. In other embodiments, fragrances may be selected based on, for example, Fig. 9 A plurality of fragrances can be selected based on the multiple differentiated decision paths described in . A fragrance can be selected based on the data of the customer being used to make predictions using a machine learning algorithm to select from the recommended fragrances. Other embodiments can select fragrances based on characteristics similar to the preferences of the customer (including similar olfaction and ingredients). In some embodiments, the fragrance recommendation can include a score indicating the possibility that the customer prefers the fragrance. A fragrance can be selected based on this score, such as the fragrance with the highest score, or when selecting multiple fragrances, the first "n" fragrances with the highest recommendation score. At step 1010, one or more selected fragrances are sent to the basic module 122.

[0054] Now refer to Fig.11 The function of the "feedback module" is explained. Those skilled in the art will understand that for this process and method and other processes and methods disclosed herein, the functions performed in the process and method may be implemented in different orders. In addition, the steps and operations outlined are provided only as examples, and some of these steps and operations may be optional, combined into fewer steps and operations, or expanded into additional steps and operations without departing from the essence of the disclosed embodiments.

[0055] This figure shows the feedback module 134. The process begins at step 1102 where a fragrance is received from the base module to select. In some embodiments, multiple fragrances can be selected. Receiving fragrances can include providing one or more fragrances to the customer, such as a portion of a sample, to allow the customer to evaluate one or more fragrances. In other embodiments, receiving fragrances can include presenting the opportunity to purchase one or more fragrances to the customer. In another embodiment, one or more fragrances can be purchased as gifts. In some embodiments, one or more fragrances can be presented to the user as a notification, such as by email, SMS message, application notification, etc. The notification or purchase opportunity can additionally include a purchase incentive, such as a discount, rebate, refund guarantee, etc. In a preferred embodiment, one or more fragrances are selected and provided according to parameters and / or recommendations customized based on the available customer characteristics of the customer. At step 1104, a feedback request is sent to the customer. The feedback request can involve the selected fragrance, or it can be a general query instead. An example of a feedback request can be a customer satisfaction survey delivered by mail or email. The feedback request is optional and is not required for practicing the present invention. Alternatively, customers can voluntarily share their feedback by such as returning the product or buying the product again or by voluntarily submitting comments, complaints, etc. At step 1106, feedback is received from the customer. Feedback can be explicit, such as from survey responses, comments, complaints, etc., or can be passively obtained by observing actions taken by customers, such as whether they purchase products containing fragrances, which would indicate a positive response or preference, particularly if one or more products with the same fragrance are purchased repeatedly, while customers returning products with a specific fragrance can indicate a negative response or preference. Customer feedback can also be obtained through social media posts, including mentions, shares, recommendations, and critical reviews. In some embodiments, a variety of fragrances can be presented to the customer and the customer provides feedback about their preferences. The feedback can include a ranking of the provided fragrances from most preferred to least preferred. The feedback can alternatively be binary, such as preferred or not preferred, liked or disliked, etc. The feedback can alternatively be an independent score, such as 8 / 10, where 10 / 10 represents the highest possible preference and 0 / 10 represents the worst or lowest possible preference. At step 1108, the customer feedback is saved to the customer database 118. The information accumulated in the customer database 118 can be used to further refine the customer feedback. Figures 6 to 10 Any activity described in for analyzing, associating, recommending or selecting fragrance. At step 1110, return to base module 122.

[0056] In a preferred embodiment, a plurality of questions are provided to the user or customer, and the customer responds to the plurality of questions via input device 110, microphone 108, camera 106, or detects the response to the plurality of questions via one or more sensors 104. In addition to questions intended to help identify the personality of the user, these questions also include identification questions. The questions may additionally request the customer to declare one or more preferences for one or more fragrances, fragrance properties, cosmetics or clothing styles, lifestyles, etc. The customer may additionally be requested to select one or more fragrances they are currently using or wish to use. The customer may also rank their fragrance preferences and / or the fragrances they are currently using in an ordinal manner. One or more images may be obtained via one or more cameras 106 or via file upload. The one or more images are analyzed to determine one or more personality traits, or alternative image methods (e.g., video clips or questionnaires) may be used to determine one or more personality traits. One or more recommended paths are identified from the information provided by the customer. Examples include personality, fragrance similarity to currently used fragrances, similarity to claimed preferences, ingredient similarity to currently used fragrances, a combination of personality and fragrance similarity to currently used fragrances using weighted values, and a combination of personality and ingredient similarity to currently used fragrances. A fragrance recommendation may be selected for at least one of the recommended paths for each recommendation. For example, the user's personality may be identified as adventurous, and a matching fragrance recommendation may be Her Eau de Parfum. The selected recommendation may be provided to the customer, feedback may be received from the customer and used to evaluate the relevance of the selected recommendation.

[0057] Fig.12 According to an embodiment, fragrance is shown as a beauty and style proposition. Fig.12 Fragrances as part of an overall beauty and style proposition for various occasions and a multifaceted fragrance language and expression are shown.

[0058] Fig.13 According to an embodiment, a component of a fragrance recommendation is shown. Fig.13 The experience component, service component, content component, and product component of the fragrance recommendation system are shown.

[0059] Fig.14 According to an embodiment, an example of a workflow for making a fragrance recommendation is shown. Fig.14 An online personal fragrance consultation workflow of a fragrance recommendation system is shown.

[0060] Fig.15 According to an embodiment, an example of a gifting workflow for making a fragrance recommendation is shown. Fig.15 The gifting workflow of the fragrance recommendation system is shown.

[0061] Fig.16 According to an embodiment, an example of a personality connection between a conversation initiator and a refinement trigger is shown. Fig.16 The personality trait and sub-trait score connections between conversation initiators and refinement triggers are shown.

[0062] Fig.17 Examples of refining trigger factors are shown according to the embodiments. Fig.17 An example of a stimulus or activity to which a user is exposed is shown when performing facial analysis to observe and determine the user's reaction to the stimulus or activity.

[0063] Fig.18 According to an embodiment, an example of customer to product mapping is shown. Fig.18 Mapping consumers to products using consumer profiles, brand archetypes, fragrance usage and attitudes, educational scent personalities, and data science iterations is shown.

[0064] Fig.19 According to an embodiment, an example of a scent-to-personality database is shown. Fig.19 An exemplary database architecture for mapping scents to personality keywords is shown.

[0065] Fig. 20A , 21 B and 21C show the classification of personality traits according to an embodiment. Figures 20A to 20C Examples of keyword groupings for the attitude, feeling, and perception personality dimensions are shown.

[0066] Fig.21 An example of keyword grouping is shown according to the embodiment. Fig.21 Exemplary personality traits mapped to keyword groupings are shown.

[0067] Fig. 22 is a flow chart of an example of a method 2200 for making fragrance recommendations according to an embodiment. The method 2200 may provide Figures 1 to 22 Features described in .

[0068] At operation 2205, sensor data is collected from a sensor array of a user computing device. In an example, the user computing device may be a mobile computing device, a smartphone, a tablet computing device, a laptop computing device, a desktop computing device, or a kiosk.

[0069] At operation 2210, user data of a user associated with a user computing device is obtained. In an example, the user data may include the user's gender, the user's hobbies, the user's personality traits, the user's fragrance preferences, the user's purchase history, or user-submitted product feedback.

[0070] At operation 2215, customer attribute features are extracted from the sensor data and the user data. In an example, the customer attribute features may include at least one human parameter and at least one non-human parameter. In an example, the sensor data is evaluated to determine the current emotional state of the user, and the current emotional state is added to the customer attribute features. In an example, an image of the user may be obtained from the sensor data, and facial features may be detected in the image. The facial features may be evaluated to determine the customer attribute features, and the customer attribute features may be added to the customer attribute features. In an example, a genetic profile of the user may be obtained. The genetic profile may be evaluated to identify the genetic customer attribute features, and the genetic customer attribute features may be added to the customer attribute features. In an example, social media data may be extracted from the user data. A social media profile may be identified using the social media data. Characteristic data may be obtained using the social media profile. Social media attribute features may be extracted from the characteristic data, and the social media attribute features may be added to the customer attribute features. In an example, environmental condition data of the user's environment may be obtained. Environmental attribute features may be extracted from the environmental condition data, and the environmental attribute features may be added to the customer attribute features. In an example, a video feed may be obtained from a sensor array. The video feed may be processed using an artificial intelligence processor to identify the user's emotional response to stimuli present in the video feed. An emotional response attribute feature of a user can be generated based on the identified emotional response, and the emotional response feature can be added to the customer attribute feature. In an example, an air sample can be collected using a volatile organic compound (VOC) sensor. The air sample can be evaluated to identify the concentration of a set of VOCs. The concentration of a set of VOCs can be used to query a fragrance database to identify the current fragrance. Fragrance attribute features of the current fragrance can be obtained from the fragrance database, and the fragrance attribute features can be added to the customer attribute features. In an example, social proximity data of a user can be obtained. Social proximity data can be evaluated to identify a connection between a user and a connection. Connection data of a connection can be collected. Connection attribute features can be extracted from the connection data, and the connection attribute features can be added to the customer attribute features. In an example, an event can be identified in the user data. User data can be evaluated to collect event data. Event attribute features can be extracted from the event data, and the event attribute features can be added to the customer attribute features.

[0071] At operation 2220, a first recommendation path is established to evaluate the customer attribute features using the fragrance recommendation model to calculate a set of customer preference probabilities of fragrances in the fragrance database. In an example, fragrance user characteristic training data is obtained, and the fragrance recommendation model is trained using the fragrance user characteristic training data to calculate the customer preference probability of fragrances based on customer characteristics. In an example, reaction data can be captured from a sensor array in response to presenting a stimulus to a user via a display. Reaction attribute features can be extracted from the reaction data, and the reaction attribute features can be evaluated in combination with the customer attribute features to calculate a set of customer preference probabilities. In an example, the customer attribute features can be evaluated to assign users to preference groups, and the set of customer preference probabilities and preference groups can be used to select fragrance recommendations. In an example, a first correlation coefficient of a first feature of a customer attribute feature can be calculated, and a second correlation coefficient of a second feature of a customer attribute feature can be calculated. It can be determined that the first correlation coefficient is outside the correlation coefficient threshold, and the second correlation coefficient is within the correlation coefficient. The second feature can be ignored, and the fragrance recommendation model can be used to evaluate the first feature.

[0072] At operation 2225, a fragrance recommendation is output to the user via a display of the user computing device based on the set of customer preference probabilities. In an example, the set of customer preference probabilities may be compared to a library of sample fragrances to select a set of fragrance samples for the user. A notification message with a recommendation notification including an identification of the set of fragrance samples may be generated for the user, and the notification message may be sent to the user's device.

[0073] In an example, a response request prompt may be sent to an output device of a user computing device along with the fragrance recommendation. A response to the response request may be received via an input device of the user computing device. A second recommendation path may be established to evaluate the fragrance database using preference parameters determined based on the response. The fragrance database may be evaluated using the preference parameters to select an alternative fragrance recommendation, and the alternative fragrance recommendation may be output to the user via a display of the user computing device.

[0074] In an example, a genetic profile of a user may be obtained, and a list of ingredients for a fragrance may be obtained. The genetic profile may be evaluated to generate a prediction of an allergic reaction to ingredients in the list of ingredients. A recommendation for the fragrance may be prevented based on the prediction of an allergic reaction, and an alternative fragrance may be selected based on the genetic profile and a set of customer preference probabilities.

[0075] In an example, a feedback request prompt may be sent to an output device of the user computing device along with the fragrance recommendation.Feedback regarding the fragrance recommendation may be received from the user via an input device of the user computing device, and the feedback may be used to refine the fragrance recommendation model.

[0076] In an example, a training data corpus including personality traits and corresponding fragrance preferences can be used to train a personality fragrance prediction model. The personality fragrance prediction model can be used to evaluate fragrance attributes of fragrances from a fragrance database to predict one or more personality traits associated with the fragrance attributes, and the predicted one or more personality traits associated with the fragrance attributes can be stored in the fragrance database.

[0077] Fig.23 A block diagram of an exemplary machine 2300 is shown, on which any one or more of the techniques (e.g., methods) discussed herein can be performed. In alternative embodiments, the machine 2300 can operate as a standalone device, or can be connected (e.g., networked) to other machines. In a networked deployment, the machine 2300 can operate in the capacity of a server computer, a client machine, or both in a server-client network environment. In an example, the machine 2300 can act as a peer machine in a peer-to-peer (P2P) (or other distributed) network environment. The machine 2300 can be a personal computer (PC), a tablet PC, a set-top box (STB), a personal digital assistant (PDA), a mobile phone, a network device, a network router, a switch or a bridge, or any machine capable of (sequentially or otherwise) executing instructions specifying the action to be taken by the machine. In addition, although only a single machine is shown, the term "machine" should also be deemed to include any collection of machines that execute a set (or multiple sets) of instructions individually or jointly to perform any one or more of the methods discussed herein, such as cloud computing, software as a service (SaaS), and other computer cluster configurations.

[0078] As described herein, examples may include logic or multiple components or mechanisms, or may be operated by logic or multiple components or mechanisms. A circuit set is a collection of circuits implemented in a tangible entity including hardware (e.g., simple circuits, gates, logic, etc.). Circuit set affiliations may be flexible over time and with underlying hardware variability. A circuit set includes components that can perform specified operations individually or in combination when operating. In an example, the hardware of a circuit set may be immutably designed to perform a specific operation (e.g., hard wiring). In an example, the hardware of a circuit set may include variably connected physical components (e.g., execution units, transistors, simple circuits, etc.), which include physical modifications (e.g., magnetically, electrically, removably placed particles of constant mass, etc.) to encode computer-readable media of instructions for specific operations. When connecting physical components, the underlying electrical properties of the hardware composition are changed, for example, from an insulator to a conductor, and vice versa. Instructions enable embedded hardware (e.g., execution units or loading mechanisms) to create components of a circuit set in hardware via variable connections to perform parts of specific operations when operating. Thus, when the apparatus is operating, the computer-readable medium is communicatively coupled to other components of the circuit set component. In an example, any of the physical components may be used in more than one component in more than one circuit set. For example, in operation, an execution unit may be used in a first circuit of a first circuit set at one point in time and reused by a second circuit in the first circuit set or by a third circuit in the second circuit set at a different time.

[0079] The machine (e.g., computer system) 2300 may include a hardware processor 2302 (e.g., a central processing unit (CPU), a graphics processing unit (GPU), a hardware processor core, or any combination thereof), a main memory 2304, and a static memory 2306, some or all of which may communicate with each other via a link (e.g., a bus) 2308. The machine 2300 may also include a display unit 2310, an alphanumeric input device 2312 (e.g., a keyboard), and a user interface (UI) navigation device 2314 (e.g., a mouse). In an example, the display unit 2310, the input device 2312, and the UI navigation device 2314 may be a touch screen display. The machine 2300 may additionally include a storage device (e.g., a drive unit) 2316, a signal generating device 2318 (e.g., a speaker), a network interface device 2320, and one or more sensors 2321, such as a global positioning system (GPS) sensor, a compass, an accelerometer, or other sensors. The machine 2300 may include an output controller 2328, such as a serial (e.g., universal serial bus (USB), parallel or other wired or wireless (e.g., infrared (IR), near field communication (NFC), etc.) connection to communicate with or control one or more peripheral devices (e.g., a printer, a card reader, etc.).

[0080] The storage device 2316 may include a machine-readable medium 2322 on which is stored one or more sets of data structures or instructions 2324 (e.g., software) embodying or used by any one or more of the techniques or functions described herein. The instructions 2324 may also reside, completely or at least partially, within the main memory 2304, within the static memory 2306, or within the hardware processor 2302 during execution thereof by the machine 2300. In an example, one or any combination of the hardware processor 2302, the main memory 2304, the static memory 2306, or the storage device 2316 may constitute a machine-readable medium.

[0081] Although the machine-readable medium 2322 is illustrated as a single medium, the term “machine-readable medium” may include a single medium or multiple media (eg, a centralized or distributed database and / or associated caches and servers) configured to store one or more instructions 2324 .

[0082] The term "machine-readable medium" may include any medium capable of storing, encoding, or carrying instructions for execution by the machine 2300 and causing the machine 2300 to perform any one or more of the techniques of the present disclosure, or any medium capable of storing, encoding, or carrying data structures used by or associated with such instructions. Non-limiting examples of machine-readable media may include solid-state memory and optical and magnetic media. In an example, a machine-readable medium may exclude a temporary propagating signal (e.g., a non-temporary machine-readable storage medium). Specific examples of non-temporary machine-readable storage media may include: non-volatile memory, such as semiconductor memory devices (e.g., electrically programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM)) and flash memory storage devices; magnetic disks, such as internal hard disks and removable disks; magneto-optical disks; and CD-ROM and DVD-ROM disks.

[0083] The instructions 2324 may also be sent or received over a communication network 2326 using a transmission medium via the network interface device 2320 using any of a variety of transmission protocols (e.g., frame relay, Internet Protocol (IP), transmission control protocol (TCP), user datagram protocol (UDP), hypertext transfer protocol (HTTP), etc.). 23 Exemplary communication networks may include a local area network (LAN), a wide area network (WAN), a packet data network (e.g., the Internet), a mobile telephone network (e.g., a cellular network), a plain old telephone (POTS) network, and a wireless data network (e.g., the Institute of Electrical and Electronics Engineers (IEEE) 802.11 family of standards, known as LPWAN standards, etc.), IEEE 802.15.4 standard series, peer-to-peer (P2P) networks, 3rd Generation Partnership Project (3GPP) standards for 4G and 5G wireless communications, including: 3GPP Long Term Evolution (LTE) standard series, 3GPP LTE Advanced standard series, 3GPP LTE Advanced Pro standard series, 3GPP New Radio (NR) standard series, etc. In an example, the network interface device 2320 may include one or more physical sockets (e.g., Ethernet, coaxial or telephone sockets) or one or more antennas to connect to the communication network 2326. In an example, the network interface device 2320 may include multiple antennas to wirelessly communicate using at least one of single-input multiple-output (SIMO), multiple-input multiple-output (MIMO), or multiple-input single-output (MISO) technology. The term "transmission medium" should be deemed to include any intangible medium capable of storing, encoding or carrying instructions for execution by the machine 2300, and includes digital or analog communication signals or other intangible media to facilitate the communication of such software.

[0084] The functions performed in these processes and methods may be implemented in different orders. In addition, the steps and operations outlined are provided only as examples, and some of these steps and operations may be optional, combined into fewer steps and operations, or expanded into additional steps and operations without departing from the essence of the disclosed embodiments.

[0085] Additional Examples

[0086] Example 1 is a method that can be executed by a computing circuit, comprising: collecting sensor data from a sensor array of a user computing device; obtaining user data of a user associated with the user computing device; extracting customer attribute characteristics from the sensor data and the user data; establishing a first recommendation path to evaluate the customer attribute characteristics using a fragrance recommendation model to calculate a set of customer preference probabilities for fragrances in a fragrance database; and based on the set of customer preference probabilities, outputting a fragrance recommendation to the user via a display of the user computing device.

[0087] In Example 2, the subject matter of Example 1 includes: obtaining fragrance user characteristic training data; and training the fragrance recommendation model using the fragrance user characteristic training data to calculate a customer preference probability of a fragrance based on customer characteristics.

[0088] In Example 3, the subject matter of Examples 1-2 includes: evaluating the sensor data to determine a current emotional state of the user; and adding the current emotional state to the customer attribute feature.

[0089] In Example 4, the subject matter of Examples 1-3, wherein the user computing device is a mobile computing device, a smartphone, a tablet computing device, a laptop computing device, a desktop computing device, or a kiosk.

[0090] In Example 5, the subject matter of Examples 1-4 includes: capturing reaction data from the sensor array in response to presenting a stimulus to the user via the display; extracting reaction attribute features from the reaction data; and evaluating the reaction attribute features in combination with the customer attribute features to calculate the set of customer preference probabilities.

[0091] In Example 6, the subject matter of Examples 1-5, wherein the user data includes the user's gender, the user's hobbies, the user's personality traits, the user's fragrance preferences, the user's purchase history, or product feedback submitted by the user.

[0092] In Example 7, the subject matter of Examples 1-6 includes: obtaining an image of the user from the sensor data; detecting facial features in the image; evaluating the facial features to determine a customer attribute feature; and adding the customer attribute feature to the customer attribute feature.

[0093] In Example 8, the subject matter of Examples 1-7 includes: comparing the set of customer preference probabilities with a sample fragrance library to select a set of fragrance samples for the user; generating a notification message for the user with a recommendation notification including an identification of the set of fragrance samples; and sending the notification message to the user's device.

[0094] In Example 9, the subject matter of Examples 1-8 includes: obtaining a genetic profile of the user; obtaining a list of ingredients for the fragrance; evaluating the genetic profile to generate a prediction of an allergic reaction to ingredients in the list of ingredients; preventing recommendation of the fragrance based on the prediction of the allergic reaction; and selecting an alternative fragrance based on the genetic profile and the set of customer preference probabilities.

[0095] In Example 10, the subject matter of Examples 1-9 includes: obtaining a genetic profile of the user; evaluating the genetic profile to identify a genetic customer attribute characteristic; and adding the genetic customer attribute characteristic to the customer attribute characteristic.

[0096] In Example 11, the subject matter of Examples 1-10 includes: extracting social media data from the user data; identifying a social media profile using the social media data; obtaining characteristic data using the social media profile; extracting social media attribute features from the characteristic data; and adding the social media attribute features to the customer attribute features.

[0097] In Example 12, the subject matter of Examples 1-11 includes: obtaining environmental condition data of the environment of the user; extracting environmental attribute features from the environmental condition data; and adding the environmental attribute features to the customer attribute features.

[0098] In Example 13, the subject matter of Examples 1-12 includes: evaluating the customer attribute characteristics to assign the user to a preference group; and selecting the fragrance recommendation using the set of customer preference probabilities and the preference group.

[0099] In Example 14, the subject matter of Examples 1-13 includes: obtaining a video feed from the sensor array; processing the video feed using an artificial intelligence processor to identify an emotional response of the user to a stimulus present in the video feed; generating an emotional response attribute feature for the user based on the identified emotional response; and adding the emotional response feature to the customer attribute feature.

[0100] In Example 15, the subject matter of Examples 1-14 includes: collecting air samples using a volatile organic compound (VOC) sensor; evaluating the air samples to identify concentrations of a set of VOCs; querying the fragrance database using the concentrations of the set of VOCs to identify a current fragrance; obtaining fragrance attribute characteristics of the current fragrance from the fragrance database; and adding the fragrance attribute characteristics to the customer attribute characteristics.

[0101] In Example 16, the subject matter of Examples 1-15 includes: obtaining social proximity data for the user; evaluating the social proximity data to identify a connection between the user and a connection; collecting connection data for the connection; extracting connection attribute features from the connection data; and adding the connection attribute features to the customer attribute features.

[0102] In Example 17, the subject matter of Examples 1-16 includes: calculating a first correlation coefficient for a first feature of the customer attribute feature and a second correlation coefficient for a second feature of the customer attribute feature; determining that the first correlation coefficient is outside a correlation coefficient threshold and the second correlation coefficient is within the correlation coefficient threshold; ignoring the second feature; and evaluating the first feature using the fragrance recommendation model.

[0103] In Example 18, the subject matter of Examples 1-17 includes: sending a response request prompt and the fragrance recommendation to an output device of the user computing device; receiving a response to the response request via an input device of the user computing device; establishing a second recommendation path to evaluate the fragrance database using preference parameters determined based on the response; evaluating the fragrance database using the preference parameters to select an alternative fragrance recommendation; and outputting the alternative fragrance recommendation to the user via a display of the user computing device.

[0104] In Example 19, the subject matter of Examples 1-18, wherein the customer attribute characteristics include at least one human parameter and at least one non-human parameter.

[0105] In Example 20, the subject matter of Examples 1-19 includes: sending a feedback request prompt along with the fragrance recommendation to an output device of the user computing device; receiving feedback from the user regarding the fragrance recommendation via an input device of the user computing device; and refining the fragrance recommendation model using the feedback.

[0106] In Example 21, the subject matter of Examples 1-20 includes: identifying events in the user data; evaluating the user data to collect event data; extracting event attribute features from the event data; and adding the event attribute features to the customer attribute features.

[0107] In Example 22, the subject matter of Examples 1-21 includes: training a personalized fragrance prediction model using a training data corpus including personality traits and corresponding fragrance preferences; using the personalized fragrance prediction model to evaluate fragrance attributes of fragrances from the fragrance database to predict one or more personality traits associated with the fragrance attributes; and storing the predicted one or more personality traits associated with the fragrance attributes in the fragrance database.

[0108] Example 23 is a system comprising components for performing any of the methods of Examples 1-22.

[0109] Example 24 is at least one machine-readable medium comprising instructions that, when executed by a machine, cause the machine to perform any of the methods of Examples 1-22.

[0110] Example 25 is a system comprising: at least one processor; and a memory comprising instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: collect sensor data from a sensor array of a user computing device; obtain user data of a user associated with the user computing device; extract customer attribute characteristics from the sensor data and the user data; establish a first recommendation path to evaluate the customer attribute characteristics using a fragrance recommendation model to calculate a set of customer preference probabilities for fragrances in a fragrance database; and based on the set of customer preference probabilities, output a fragrance recommendation to the user via a display of the user computing device.

[0111] In Example 26, the subject matter of Example 25 includes that the memory further includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: obtain fragrance user characteristic training data; and use the fragrance user characteristic training data to train the fragrance recommendation model to calculate a customer preference probability of a fragrance based on customer characteristics.

[0112] In Example 27, the subject matter of Examples 25-26 includes that the memory further includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: evaluate the sensor data to determine a current emotional state of the user; and add the current emotional state to the customer attribute feature.

[0113] In Example 28, the subject matter of Examples 25-27, wherein the user computing device is a mobile computing device, a smartphone, a tablet computing device, a laptop computing device, a desktop computing device, or a kiosk.

[0114] In Example 29, the subject matter of Examples 25-28 includes that the memory also includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: capture reaction data from the sensor array in response to presenting stimuli to the user via the display; extract reaction attribute features from the reaction data; and evaluate the reaction attribute features in combination with the customer attribute features to calculate the set of customer preference probabilities.

[0115] In Example 30, the subject matter of Examples 25-29, wherein the user data includes the user's gender, the user's hobbies, the user's personality traits, the user's fragrance preferences, the user's purchase history, or product feedback submitted by the user.

[0116] In Example 31, the subject matter of Examples 25-30 includes that the memory also includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: obtain an image of the user from the sensor data; detect facial features in the image; evaluate the facial features to determine customer attribute features; and add the customer attribute features to the customer attribute features.

[0117] In Example 32, the subject matter of Examples 25-31 includes that the memory also includes instructions, which when executed by the at least one processor cause the at least one processor to perform operations to: compare the set of customer preference probabilities with the sample fragrance library to select a set of fragrance samples for the user; generate a notification message for the user with a recommendation notification including an identification of the set of fragrance samples; and send the notification message to the user's device.

[0118] In Example 33, the subject matter of Examples 25-32 includes that the memory also includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: obtain a genetic profile of the user; obtain a list of ingredients for the fragrance; evaluate the genetic profile to generate a prediction of an allergic reaction to ingredients in the list of ingredients; prevent recommendation of the fragrance based on the prediction of the allergic reaction; and select an alternative fragrance based on the genetic profile and the set of customer preference probabilities.

[0119] In Example 34, the subject matter of Examples 25-33 includes that the memory also includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: obtain a genetic profile of the user; evaluate the genetic profile to identify a genetic customer attribute characteristic; and add the genetic customer attribute characteristic to the customer attribute characteristic.

[0120] In Example 35, the subject matter of Examples 25-34 includes that the memory also includes instructions, which, when executed by the at least one processor, cause the at least one processor to perform operations to: extract social media data from the user data; identify a social media profile using the social media data; obtain characteristic data using the social media profile; extract social media attribute features from the characteristic data; and add the social media attribute features to the customer attribute features.

[0121] In Example 36, the subject matter of Examples 25-35 includes that the memory also includes instructions, which when executed by the at least one processor cause the at least one processor to perform operations to: obtain environmental condition data of the user's environment; extract environmental attribute characteristics from the environmental condition data; and add the environmental attribute characteristics to the customer attribute characteristics.

[0122] In Example 37, the subject matter of Examples 25-36 includes that the memory further includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: evaluate the customer attribute characteristics to assign the user to a preference group; and select the fragrance recommendation using the set of customer preference probabilities and the preference group.

[0123] In Example 38, the subject matter of Examples 25-37 includes, the memory further comprising instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: obtain a video feed from the sensor array; process the video feed using an artificial intelligence processor to identify an emotional response of the user to stimuli present in the video feed; generate an emotional response attribute feature for the user based on the identified emotional response; and add the emotional response feature to the customer attribute feature.

[0124] In Example 39, the subject matter of Examples 25-38 includes that the memory also includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: collect air samples using a volatile organic compound (VOC) sensor; evaluate the air samples to identify concentrations of a set of VOCs; query the fragrance database using the concentrations of the set of VOCs to identify a current fragrance; obtain fragrance attribute characteristics of the current fragrance from the fragrance database; and add the fragrance attribute characteristics to the customer attribute characteristics.

[0125] In Example 40, the subject matter of Examples 25-39 includes that the memory also includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: obtain social proximity data of the user; evaluate the social proximity data to identify a connection between the user and a connection; collect connection data of the connection; extract connection attribute features from the connection data; and add the connection attribute features to the customer attribute features.

[0126] In Example 41, the subject matter of Examples 25-40 includes that the memory also includes instructions, which when executed by the at least one processor cause the at least one processor to perform operations to: calculate a first correlation coefficient of a first feature of the customer attribute characteristics and a second correlation coefficient of a second feature of the customer attribute characteristics; determine that the first correlation coefficient is outside a correlation coefficient threshold and the second correlation coefficient is within the correlation coefficient threshold; ignore the second feature; and evaluate the first feature using the fragrance recommendation model.

[0127] In Example 42, the subject matter of Examples 25-41 includes that the memory also includes instructions, which when executed by the at least one processor cause the at least one processor to perform operations to: send a response request prompt and the fragrance recommendation to an output device of the user computing device; receive a response to the response request via an input device of the user computing device; establish a second recommendation path to evaluate the fragrance database using preference parameters determined based on the response; evaluate the fragrance database using the preference parameters to select an alternative fragrance recommendation; and output the alternative fragrance recommendation to the user via a display of the user computing device.

[0128] In Example 43, the subject matter of Examples 25-42, wherein the customer attribute characteristics include at least one human parameter and at least one non-human parameter.

[0129] In Example 44, the subject matter of Examples 25-43 includes that the memory also includes instructions, which when executed by the at least one processor cause the at least one processor to perform operations to: send a feedback request prompt along with the fragrance recommendation to an output device of the user computing device; receive feedback about the fragrance recommendation from the user via an input device of the user computing device; and use the feedback to refine the fragrance recommendation model.

[0130] In Example 45, the subject matter of Examples 25-44 includes that the memory also includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: identify events in the user data; evaluate the user data to collect event data; extract event attribute features from the event data; and add the event attribute features to the customer attribute features.

[0131] In Example 46, the subject matter of Examples 25-45 includes that the memory also includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: train a personalized fragrance prediction model using a training data corpus including personality traits and corresponding fragrance preferences; use the personalized fragrance prediction model to evaluate fragrance attributes of fragrances from the fragrance database to predict one or more personality traits associated with the fragrance attributes; and store the predicted one or more personality traits associated with the fragrance attributes in the fragrance database.

[0132] Example 47 is at least one non-transitory machine-readable medium, comprising instructions that, when executed by at least one processor, cause the at least one processor to perform operations to: collect sensor data from a sensor array of a user computing device; obtain user data of a user associated with the user computing device; extract customer attribute characteristics from the sensor data and the user data; establish a first recommendation path to evaluate the customer attribute characteristics using a fragrance recommendation model to calculate a set of customer preference probabilities for fragrances in a fragrance database; and based on the set of customer preference probabilities, output a fragrance recommendation to the user via a display of the user computing device.

[0133] In Example 48, the subject matter of Example 47 includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: obtain fragrance user characteristic training data; and train the fragrance recommendation model using the fragrance user characteristic training data to calculate a customer preference probability of a fragrance based on customer characteristics.

[0134] In Example 49, the subject matter of Examples 47-48 includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: evaluate the sensor data to determine a current emotional state of the user; and add the current emotional state to the customer attribute feature.

[0135] In Example 50, the subject matter of Examples 47-49, wherein the user computing device is a mobile computing device, a smartphone, a tablet computing device, a laptop computing device, a desktop computing device, or a kiosk.

[0136] In Example 51, the subject matter of Examples 47-50 includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: capture reaction data from the sensor array in response to presenting stimuli to the user via the display; extract reaction attribute features from the reaction data; and evaluate the reaction attribute features in combination with the customer attribute features to calculate the set of customer preference probabilities.

[0137] In Example 52, the subject matter of Examples 47-51, wherein the user data includes the user's gender, the user's hobbies, the user's personality traits, the user's fragrance preferences, the user's purchase history, or product feedback submitted by the user.

[0138] In Example 53, the subject matter of Examples 47-52 includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: obtain an image of the user from the sensor data; detect facial features in the image; evaluate the facial features to determine customer attribute features; and add the customer attribute features to the customer attribute features.

[0139] In Example 54, the subject matter of Examples 47-53 includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: compare the set of customer preference probabilities with a sample fragrance library to select a set of fragrance samples for the user; generate a notification message for the user with a recommendation notification including an identification of the set of fragrance samples; and send the notification message to the user's device.

[0140] In Example 55, the subject matter of Examples 47-54 includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: obtain a genetic profile of the user; obtain a list of ingredients for the fragrance; evaluate the genetic profile to generate a prediction of an allergic reaction to ingredients in the list of ingredients; prevent recommendation of the fragrance based on the prediction of the allergic reaction; and select an alternative fragrance based on the genetic profile and the set of customer preference probabilities.

[0141] In Example 56, the subject matter of Examples 47-55 includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: obtain a genetic profile of the user; evaluate the genetic profile to identify a genetic customer attribute characteristic; and add the genetic customer attribute characteristic to the customer attribute characteristic.

[0142] In Example 57, the subject matter of Examples 47-56 includes instructions that, when executed by the at least one processor, cause the at least one processor to perform the following operations: extract social media data from the user data; identify a social media profile using the social media data; obtain characteristic data using the social media profile; extract social media attribute features from the characteristic data; and add the social media attribute features to the customer attribute features.

[0143] In Example 58, the subject matter of Examples 47-57 includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: obtain environmental condition data of the user's environment; extract environmental attribute characteristics from the environmental condition data; and add the environmental attribute characteristics to the customer attribute characteristics.

[0144] In Example 59, the subject matter of Examples 47-58 includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: evaluate the customer attribute characteristics to assign the user to a preference group; and select a fragrance recommendation using the set of customer preference probabilities and the preference group.

[0145] In Example 60, the subject matter of Examples 47-59 includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: obtain a video feed from the sensor array; process the video feed using an artificial intelligence processor to identify an emotional response of the user to stimuli present in the video feed; generate an emotional response attribute feature for the user based on the identified emotional response; and add the emotional response feature to the customer attribute feature.

[0146] In Example 61, the subject matter of Examples 47-60 includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: collect an air sample using a volatile organic compound (VOC) sensor; evaluate the air sample to identify a set of VOC concentrations; query the fragrance database using the set of VOC concentrations to identify a current fragrance; obtain fragrance attribute characteristics of the current fragrance from the fragrance database; and add the fragrance attribute characteristics to the customer attribute characteristics.

[0147] In Example 62, the subject matter of Examples 47-61 includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: obtain social proximity data for the user; evaluate the social proximity data to identify a connection between the user and a connection; collect connection data for the connection; extract connection attribute features from the connection data; and add the connection attribute features to the customer attribute features.

[0148] In Example 63, the subject matter of Examples 47-62 includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: calculate a first correlation coefficient for a first feature of the customer attribute characteristics and a second correlation coefficient for a second feature of the customer attribute characteristics; determine that the first correlation coefficient is outside a correlation coefficient threshold and that the second correlation coefficient is within the correlation coefficient threshold; ignore the second feature; and evaluate the first feature using the fragrance recommendation model.

[0149] In Example 64, the subject matter of Examples 47-63 includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: send a response request prompt along with the fragrance recommendation to an output device of the user computing device; receive a response to the response request via an input device of the user computing device; establish a second recommendation path to evaluate the fragrance database using preference parameters determined based on the response; evaluate the fragrance database using the preference parameters to select an alternative fragrance recommendation; and output the alternative fragrance recommendation to the user via a display of the user computing device.

[0150] In Example 65, the subject matter of Examples 47-64, wherein the customer attribute characteristics include at least one human parameter and at least one non-human parameter.

[0151] In Example 66, the subject matter of Examples 47-65 includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: send a feedback request prompt along with the fragrance recommendation to an output device of the user computing device; receive feedback about the fragrance recommendation from the user via an input device of the user computing device; and use the feedback to refine the fragrance recommendation model.

[0152] In Example 67, the subject matter of Examples 47-66 includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: identify events in the user data; evaluate the user data to collect event data; extract event attribute features from the event data; and add the event attribute features to the customer attribute features.

[0153] In Example 68, the subject matter of Examples 47-67 includes instructions that, when executed by the at least one processor, cause the at least one processor to perform operations to: train a personalized fragrance prediction model using a training data corpus comprising personality traits and corresponding fragrance preferences; evaluate fragrance attributes of fragrances from the fragrance database using the personalized fragrance prediction model to predict one or more personality traits associated with the fragrance attributes; and store the predicted one or more personality traits associated with the fragrance attributes in the fragrance database.

[0154] Example 69 is at least one machine-readable medium comprising instructions that, when executed by processing circuitry, cause the processing circuitry to perform operations to implement any of Examples 1-68.

[0155] Example 70 is a component comprising a module to implement any of Examples 1-68.

[0156] Example 71 is a system for implementing any of Examples 1-68.

[0157] Example 72 is a method for implementing any of Examples 1-68.

Claims

1. A method executable by a computing circuit system, include: collecting sensor data from a sensor array of a user computing device; obtaining user data of a user associated with the user computing device; extracting customer attribute features from the sensor data and the user data; establishing a first recommendation path to evaluate the customer attribute characteristics using a fragrance recommendation model to calculate a set of customer preference probabilities for fragrances in a fragrance database; as well as Based on the set of customer preference probabilities, a fragrance recommendation is output to the user via a display of the user computing device.

2. The method according to claim 1, further comprising: include: Obtaining fragrance user characteristic training data; as well as The fragrance recommendation model is trained using the fragrance user characteristic training data to calculate customer preference probabilities for fragrances based on customer characteristics.

3. The method according to claim 1, further comprising: include: evaluating the sensor data to determine a current emotional state of the user; as well as The current emotional state is added to the customer attribute feature.

4. The method according to claim 1, in, The user computing device is a mobile computing device, a smart phone, a tablet computing device, a laptop computing device, a desktop computing device, or a kiosk.

5. The method according to claim 1, further comprising: include: capturing reaction data from the sensor array in response to presenting a stimulus to the user via the display; extracting reaction attribute features from the reaction data; as well as The reaction attribute characteristics are evaluated in combination with the customer attribute characteristics to calculate the set of customer preference probabilities.

6. The method according to claim 1, in, The user data includes the user's gender, the user's hobbies, the user's personality traits, the user's fragrance preferences, the user's purchase history, or product feedback submitted by the user.

7. The method according to claim 1, further comprising: include: obtaining an image of the user from the sensor data; detecting facial features in the image; evaluating said facial features to determine customer attribute characteristics; as well as The customer attribute characteristics are added to the customer attribute characteristics.

8. The method according to claim 1, further comprising: include: comparing the set of customer preference probabilities to a library of sample fragrances to select a set of fragrance samples for the user; generating a notification message for the user having a recommendation notification including an identification of the set of fragrance samples; as well as The notification message is sent to a device of the user.

9. The method according to claim 1, further comprising: include: obtaining a genetic profile of the user; Obtaining a list of ingredients of the fragrance; evaluating the genetic profile to generate a prediction of an allergic reaction to an ingredient in the list of ingredients; preventing recommendation of said fragrance based on prediction of said allergic reaction; as well as A replacement fragrance is selected based on the genetic profile and the set of customer preference probabilities.

10. The method according to claim 1, further comprising: include: obtaining a genetic profile of the user; evaluating the genetic profile to identify genetic customer attribute characteristics; as well as The inherited customer attribute characteristics are added to the customer attribute characteristics.

11. The method according to claim 1, further comprising: include: extracting social media data from the user data; identifying a social media profile using the social media data; obtaining characteristic data using the social media profile; extracting social media attribute features from the characteristic data; as well as The social media attribute features are added to the customer attribute features.

12. The method according to claim 1, further comprising: include: obtaining environmental condition data of the user's environment; extracting environmental attribute features from the environmental condition data; as well as The environmental attribute features are added to the customer attribute features.

13. The method according to claim 1, further comprising: include: evaluating the customer attribute characteristics to assign the user to a preference group; as well as The fragrance recommendation is selected using the set of customer preference probabilities and the preference set.

14. The method according to claim 1, further comprising: include: obtaining a video feed from the sensor array; processing the video feed using an artificial intelligence processor to identify an emotional response of the user to a stimulus present in the video feed; generating an emotional response attribute feature for the user based on the identified emotional response; and The emotional response feature is added to the customer attribute feature.

15. The method according to claim 1, further comprising: include: Collect air samples using a volatile organic compound (VOC) sensor; evaluating the air sample to identify concentrations of a set of VOCs; querying the fragrance database using the set of VOC concentrations to identify a current fragrance; Obtaining a fragrance attribute characteristic of the current fragrance from the fragrance database; as well as The fragrance attribute profile is added to the customer attribute profile.

16. The method according to claim 1, further comprising: include: obtaining social proximity data of the user; evaluating the social proximity data to identify a connection between the user and a connection; collecting connection data of the connection; extracting connection attribute features from the connection data; as well as The connection attribute characteristics are added to the customer attribute characteristics.

17. The method according to claim 1, further comprising: include: Calculating a first correlation coefficient of a first feature of the customer attribute feature and a second correlation coefficient of a second feature of the customer attribute feature; determining that the first correlation coefficient is outside a correlation coefficient threshold and the second correlation coefficient is within the correlation coefficient threshold; Ignore the second feature; and The first characteristic is evaluated using the fragrance recommendation model.

18. The method according to claim 1, further comprising: include: sending a response request prompt and the fragrance recommendation to an output device of the user computing device; receiving a response to the response request via an input device of the user computing device; establishing a second recommendation pathway to evaluate the fragrance database using preference parameters determined from the responses; evaluating the fragrance database using the preference parameters to select an alternative fragrance recommendation; as well as The alternative fragrance recommendation is output to the user via a display of the user computing device.

19. The method according to claim 1, in, The customer attribute characteristics include at least one human parameter and at least one non-human parameter.

20. The method according to claim 1, further comprising: include: sending a feedback request prompt and the fragrance recommendation to an output device of the user computing device; receiving feedback from the user regarding the fragrance recommendation via an input device of the user computing device; as well as The feedback is used to refine the fragrance recommendation model.

21. The method according to claim 1, further comprising: include: identifying events in the user data; evaluating the user data to collect event data; Extracting event attribute features from the event data; as well as The event attribute features are added to the customer attribute features.

22. The method according to claim 1, further comprising: include: training a personality fragrance prediction model using a training data corpus including personality traits and corresponding fragrance preferences; evaluating fragrance attributes of fragrances from the fragrance database using the personalized fragrance prediction model to predict one or more personality traits associated with the fragrance attributes; as well as The predicted one or more personality traits associated with the fragrance attributes are stored in the fragrance database.

23. A system comprising components for performing any of the methods according to claims 1-22.

24. At least one machine-readable medium comprising instructions which, when executed by a machine, cause the machine to perform any of the methods of claims 1-22.

Citation Information

Cited By

  • AI personal fragrance consultation and fragrance selection / recommendation

    US12705663B2