Method and system for predicting behavioral characteristics of customers in physical stores

Patent metadata

Field	Value
Publication number	US-10402837-B2
Application number	US-201615335495-A
Country	US
Kind code	B2
Filing date	Oct 27, 2016
Priority date	Oct 27, 2016
Publication date	Sep 3, 2019
Grant date	Sep 3, 2019

Abstract

The disclosed embodiments illustrate methods and systems for processing customer data to predict behavioral characteristics of a customer in a physical store. The method includes receiving customer data, pertaining to a shopping episode, of the customer from one or more sensing devices. The method further includes segmenting the shopping episode into one or more segments of time durations. The method further includes categorizing each of the one or more segments of time durations into one of an in-aisle category and a non-aisle category. The method further includes identifying hand-related actions of the customer based on the received customer data. The method further includes determining a likelihood of occurrence of an action-related category based on at least the identified one or more hand-related actions. The method further includes predicting the behavioral characteristics of the customer based on at least the determined likelihood of occurrence of the action-related category.

First claim

What is claimed is:

1. A method comprising:
detecting, using one or more sensing devices provided in one or more portable electronic devices carried by a customer, behavior data relating to a behavior of a customer when the customer is shopping in a physical store, the behavior data comprising one or more shopping-related actions in a shopping episode;
receiving, by a transceiver at a computing server, the behavior data from the one or more portable electronic devices;
identifying, by a processor of the one or more portable electronic devices, the one or more shopping-related actions based on at least one of an interaction of the customer with one or more products or services in the physical store, a location of the customer in the physical store, and a speed or walking pattern of the customer in the physical store;
segmenting, by a processor of the computing server, said shopping episode into one or more segments of time durations based on said one or more shopping-related actions;
categorizing, by said processor, each of said one or more segments of time durations into one of an in-aisle category and a non-aisle category based on at least an inter-step time duration of said customer during each of said one or more segments of time durations, wherein categorizing comprises:
generating a first graphic distribution of the shopping episode based on the inter-step time duration;
identifying one or more peak-point values of one or more peaks and one or more foot-point values of one or more foots in the first graphic distribution using a peak and valley detection algorithm; and
performing a change point detection analysis using a classifier to identify the in-aisle category and the non-aisle category, the change point detection analysis being based on a change to a classification probability as a size of a window including at least one peak or at least one valleys changes in the first graphic distribution; and

for each segment of time duration that corresponds to said in-aisle category:
identifying, by said processor, one or more hand-related actions of said customer based on said received behavior data associated with said each segment of time duration, wherein identifying the hand-related actions comprises:
generating a second graphic distribution based on gyroscope data extracted from the behavior data;
identifying a peak in the second graphical distribution;
computing features in a window corresponding to a width of the peak at half-maximum; and
computing a probability that an action is a hand action based on a combination of the gyroscope data and accelerometer data retrieved from the behavior data;

determining, by said processor, a likelihood of occurrence of an action-related category based on at least said identified one or more hand-related actions, wherein determining the likelihood of occurrence comprises:
determining a probability based on a set of features detected by the one or more sending devices and included in the behavior data, the set of features comprising at least one parameter of a gyroscope extracted from the gyroscope data and at least one parameter of an accelerometer extracted from the accelerometer data;

predicting, by said processor, behavioral characteristics of said customer based on at least said determined likelihood of occurrence of said action-related category associated with said each segment of time duration; and
rendering, by the processor, a user interface on a screen of the one or more portable electronic devices carried by the customer to display a targeted content to the customer, the targeted content being determined based on the predicted behavioral characteristics of the customer and comprising at least one of an offer, a discount, an advertisement, and a promotion about the one or more products or services associated with the physical store.

2. The method of claim 1 , wherein said behavior data comprises at least movement and gesture data of said customer during said shopping episode, wherein said shopping episode corresponds to a time duration spent by said customer in said physical store.

3. The method of claim 1 , wherein said one or more portable electronic devices correspond to at least one of a smartphone, a smart watch, and a smart band, and wherein said one or more sensing devices in said one or more portable electronic devices correspond to at least one or more of an accelerometer, a gyroscope, a magnetometer, a step counter, a light sensor, an audio sensor, a heart rate measuring device, and a battery temperature measuring device.

4. The method of claim 1 , wherein each of said one or more segments of time durations corresponds to at least one movement or gesture activity of said customer in said physical store.

5. The method of claim 1 , wherein a time duration of each of said in-aisle category and said non-aisle category is determined, by said processor, based on said one or more peak-point values and said one or more foot-point values associated with each of said one or more segments of time durations.

6. The method of claim 1 , wherein said action-related category corresponds to at least one of a first category, a second category, and a third category.

7. The method of claim 6 , wherein said first category comprises said one or more hand-related actions of said customer that correspond to at least a picking of one or more products from their corresponding locations in said physical store.

8. The method of claim 7 , wherein said second category comprises said one or more hand-related actions of said customer that correspond to at least putting back said picked one or more products to their corresponding locations in said physical store.

9. The method of claim 7 , wherein said third category comprises said one or more hand-related actions of said customer that correspond to at least putting said picked one or more products in a shopping trolley.

10. The method of claim 1 , wherein said behavioral characteristics of said customer correspond to at least a familiarity state or a knowledge state of said customer about one or more products or services based on at least time taken by said customer to check said one or more products when picked by said customer.

11. The method of claim 1 further comprising predicting, by said processor, said behavioral characteristics of said customer during said shopping episode based on at least an aggregation of said predicted behavioral characteristics of said customer during each of said one or more segments of time durations.

12. The method of claim 1 further comprising predicting, by said processor, said behavioral characteristics of said customer after a pre-defined time interval based on at least an aggregation of said predicted behavioral characteristics of said customer during each of a plurality of shopping episodes that corresponds to said pre-defined time interval, and wherein each of said plurality of shopping episodes is associated with one or more physical stores.

13. The method of claim 1 , wherein said predicted behavioral characteristics correspond to at least one of said one or more segments of time durations, said shopping episode, and a pre-defined time interval comprising a plurality of shopping episodes.

14. The method of claim 1 further comprising recommending, by said processor, a shop assistant to said customer based on at least said predicted behavioral characteristics of said customer, wherein said predicted behavioral characteristics correspond to at least one of said one or more segments of time durations, said shopping episode, and a pre-defined time interval comprising a plurality of shopping episodes.

Assignees

• Conduent Business Services Llc
• Singapore Management Univ
• Conduent Busness System Llc

Inventors

• Eswaran Sharanya
• Chander Deepthi
• Radhakrishnan Meeralakshmi
• Misra Archan
• Dasgupta Koustuv

Classifications

• G06Q30/0203

  Market surveys; Market polls · CPC title

• G06Q30/0201

  Market modelling; Market analysis; Collecting market data · CPC title

• G06Q30/0202 Primary

  Market predictions or forecasting for commercial activities · CPC title

• G06Q30/0204

  Market segmentation · CPC title