Neural network systems and methods for event parameter determination

Patent metadata

Field	Value
Publication number	US-11663448-B2
Application number	US-201916456807-A
Country	US
Kind code	B2
Filing date	Jun 28, 2019
Priority date	Jun 28, 2019
Publication date	May 30, 2023
Grant date	May 30, 2023

Abstract

Systems, methods, and computer program products are provided for determining an event parameter are provided. Event data can be matched to a grid comprising gridlines and cells defined by the gridlines. The grid can be mapped to a predetermined area. Each cell can comprise a number of events per predetermined time interval. The cells can be sorted into classes based on the number of events occurring during the predetermined time interval to produce a classified data set. Features can be extracted from the classified data set. The extracted features can be processed using a classifier to determine the event parameter for a future time interval in at least one cell of the cells, for example, crime events. The classifier can comprise a neural network. Systems can comprise one or more of a processor, a neural network, and a user interface.

First claim

What is claimed is:

1. A neural network system, the system comprising:
a processor configured to:
match event data, the event data comprising event location data and event time data, to a grid comprising gridlines and cells defined by the gridlines, wherein the grid is mapped to a predetermined area, and each cell comprises a number of events per predetermined time interval,
sort the cells into classes based on the number of events occurring during the predetermined time interval in each cell to produce a classified data set, wherein the classes are percentiles or weighted percentiles, and
extracting features from the classified data set;

a neural network configured to determine an event parameter for a future time interval in at least one cell of the cells; and
a user interface configured to enable a user to operate the neural network and the processor, and configured to display the event parameter on a map comprising the grid and the cell.

2. The system of claim 1 , wherein the features comprise a temporal feature, a spatial feature, or a spread feature, or any combination thereof.

3. The system of claim 1 , wherein the extracting comprises use of density-based spatial clustering of applications with noise (DBSCAN).

4. The system of claim 1 , wherein the neural network comprises a convolutional long short-term memory (Conv-LSTM) neural network.

5. The system of claim 1 , wherein the event comprises a crime event, sales event, a taxable event, an epidemiological event, a meteorological event, a vehicular traffic event, an internet traffic event, a utility consumption event, or a political event, or any combination thereof.

Assignees

• Conduent Business Services Llc

Inventors

• Kumar Niraj

Classifications

• G06N3/09

  Supervised learning · CPC title

• G06N3/0442

  characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/049 Primary

  Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs · CPC title

• G06N3/044

  Recurrent networks, e.g. Hopfield networks · CPC title