Attention Mechanism Order Prediction for Event Data

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods fail to accurately predict orders from data collected at events such as exhibitions, leading to inefficiencies in sales activities.

Innovation Solution

An order prediction device and method utilizing an attention mechanism to analyze event data, including customer information, action history, and questionnaire responses, to predict order probability and degree of interest, thereby enhancing prediction accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional methods are used to collect event data, then data collection is simple, but order prediction accuracy is low

Engineering Contradiction:

Improveorder prediction accuracyVSAvoidprediction system complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

An attention mechanism serves as an intermediary between event data and order prediction, automatically identifying and weighting important data items. This mediator processes the complex relationships between multiple event data items (customer information, action history, questionnaire responses) and generates accurate order predictions without requiring manual feature engineering

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces manual analysis and conventional statistical methods with a machine learning-based attention mechanism. This substitution automates the complex process of identifying important data items and their relationships, achieving high prediction accuracy through algorithmic processing rather than mechanical or manual analysis

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If all event data items are analyzed equally, then processing is simple, but important information is missed

Engineering Contradiction:

Improveprediction accuracyVSAvoiddata analysis automation

Core Design Contradiction:

Measurement precisionVSExtent of automation

Solution Approach 1:

The attention mechanism applies different weights (local quality) to different data items based on their importance. Instead of treating all event data uniformly, the system dynamically assigns higher attention to critical information (such as specific action history items or questionnaire responses) and lower attention to less relevant data, enabling precise prediction through differentiated analysis

Inventive Principle:

Principle #3Local quality

3Productivity

If manual analysis of event data is performed, then system complexity is low, but time consumption is high

Engineering Contradiction:

Improveprediction speedVSAvoidsystem complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The attention mechanism enables the system to automatically identify important data items and generate predictions without human intervention. The model self-adjusts weights and makes predictions autonomously based on input event data, eliminating the need for manual analysis while maintaining high speed and accuracy

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS20240127081A1Order prediction device, order prediction method, learning device, learning method, and recording medium

Publication Date: 2024.04.18 NEC CORP

AI summary

In an order prediction device, an acquisition means acquires event data related to an event. A prediction means predicts a degree of interest with respect to each data item included in the event data and an order prediction by using an attention mechanism based on the event data. An output means outputs, as a prediction result, the degree of interest and an order probability.