# US 9,436,906 B2 · Discovering user-behavior from trajectory-as-polygon (TaP) > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 9,436,906 B2 - Document ID: US9436906B2 - Status: Granted - Kind: B2 - Application number: 14/273,551 - Filed: 2014-05-09 - Published: 2015-11-12 - Granted: 2016-09-06 - Earliest claimed: 2014-05-09 - Assignees: Sap SE - Applicants: Ting Wang - Inventors: Ting Wang - Web page: https://patentgrep.com/patent/US9436906B2 ## CPC - G06N 3/006: based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO] - G06N 5/04: Inference or reasoning models - G06N 20/00: Machine learning - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract Disclosed is a system for analyzing user-behavior using a TaP algorithm. For example, raw data are collected and segmented to become segmented data. In this example, the TaP algorithm in combination with a sliding time window is implemented to derive a convex hull polygon. A determination of geometric properties of the derived convex hull polygon facilitates the analysis of the user-behavior. ## Claims ### Claim 1 A method of discovering user-behavior, the method comprising: collecting raw data; representing the raw data as trajectory data; segmenting the trajectory data based upon a sliding time window; deriving a convex hull polygon based upon the segmented data; determining geometric properties of the derived convex hull polygon; and generating a user-behavior analysis, in response to the determined geometric properties. ### Claim 2 (depends on 1) The method according to claim 1, wherein the collecting the raw data comprises storing locational data points from at least one user or object. ### Claim 3 (depends on 1) The method according to claim 1, wherein the representing comprises connecting a trajectory line between locational data points of the raw data, the trajectory line follows a time sequence of each locational data point. ### Claim 4 (depends on 1) The method according to claim 1, wherein the segmenting clusters the trajectory data into different segments corresponding to each time window in the sliding time window. ### Claim 5 (depends on 4) The method according to claim 4, wherein the sliding time window has a variable window size “τ”. ### Claim 6 (depends on 1) The method according to claim 1, wherein the derived convex hull polygon is derived using a convex hull algorithm in combination with the sliding time window. ### Claim 7 (depends on 1) The method according to claim 1, wherein the geometric properties define an active area, number of edges, edges length or deviation, and vertices of the derived convex hull polygon. ### Claim 8 (depends on 7) The method according to claim 7, wherein the active area indicates a user to be travelling or staying at a particular location within the sliding time window. ### Claim 9 (depends on 1) The method according to claim 1, wherein the determining geometric properties is based upon a series of derived convex hull polygons at different sliding time windows. ### Claim 10 One or more non-transitory computer-readable media storing processor-executable instructions that when executed cause one or more processors to perform operations that implement a Trajectory-as-Polygon (TaP) algorithm to discover a user-behavior comprising: collecting locational data points; representing the locational data as trajectory data; segmenting the trajectory data based upon a sliding time window; deriving a polygon based upon the segmented data; determining geometric properties of the derived polygon; and generating a user-behavior analysis, in response to the determined geometric properties. ### Claim 11 (depends on 10) The one or more computer-readable media according to claim 10, wherein the locational data points are from at least one user or object, and are defined in spatial-time domain. ### Claim 12 (depends on 10) The one or more computer-readable media according to claim 10, wherein the representing comprises connecting a trajectory line between the locational data points, the trajectory line follows a time sequence of each locational data point. ### Claim 13 (depends on 10) The one or more computer-readable media according to claim 10, wherein the segmenting clusters the trajectory data into different segments corresponding to each time window in the sliding time window. ### Claim 14 (depends on 13) The one or more computer-readable media according to claim 13, wherein the sliding time window has a variable window size “τ”. ### Claim 15 (depends on 10) The one or more computer-readable media according to claim 10, wherein the derived polygon is a convex hull polygon. ### Claim 16 (depends on 15) The one or more computer-readable media according to claim 15, wherein the deriving is implemented through a convex hull algorithm in combination with the sliding time window. ### Claim 17 (depends on 10) The one or more computer-readable media according to claim 10, wherein the geometric properties define an active area, number of edges, edges length or deviation, and vertices of the derived convex hull polygon. ### Claim 18 A system that implements a Trajectory-as-Polygon (TaP) algorithm to discover a user-behavior, the system comprising: a signal collecting system that collects raw data; and a processor coupled to the signal collecting system, the processor is configured to perform a method comprising: representing the raw data as trajectory data, segmenting the trajectory data based upon a sliding time window, deriving a convex hull polygon based upon the segmented data, determining geometric properties of the derived convex hull polygon, and generating a user-behavior analysis, in response to the determined geometric properties. ### Claim 19 (depends on 18) The system according to claim 18, wherein the segmenting clusters the trajectory data into different segments corresponding to each time window in the sliding time window. ### Claim 20 (depends on 18) The system according to claim 18, wherein the deriving is implemented through a convex hull algorithm in combination with the sliding time window.