Data Filtering for Predictive Analytics Accuracy

Solution Overview

Problem

Predictive analytics algorithms consume significant time and computing resources, especially when dealing with large amounts of input data, leading to inefficient and resource-intensive prediction processes.

Innovation Solution

A system that includes a data filter dictionary and a data filtering and noise reduction module using a latent semantic indexing algorithm to preprocess input data, creating a subset for predictive analytics, and an intelligent loop-back mechanism to dynamically update the dictionary based on prediction results, thereby optimizing data filtering and noise reduction.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If predictive analytics algorithms process large amounts of input data, then prediction accuracy is improved, but computing resources consumption and processing time increase significantly

Engineering Contradiction:

Improveprediction accuracyVSAvoidcomputing resources consumption

Core Design Contradiction:

Measurement precisionVSUse of energy by stationary object

Solution Approach 1:

The patent extracts and removes non-impactful data from the input dataset using a data filtering module that identifies and eliminates records with minimal predictive value. This extraction process creates a reduced dataset that maintains prediction accuracy while significantly lowering computing resource requirements for the predictive analytics algorithm.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies preliminary data filtering and noise reduction actions before the predictive analytics algorithm processes the data. By pre-processing the data to remove irrelevant information and reduce dimensionality, the system prepares an optimized dataset that reduces the computational burden during the actual prediction process.

Inventive Principle:

Principle #10Preliminary action

2Measurement precision

If predictive analytics algorithms process large amounts of input data, then prediction accuracy is improved, but processing time increases significantly

Engineering Contradiction:

Improveprediction accuracyVSAvoidprocessing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent extracts and removes non-impactful data from the input dataset using a data filtering module that identifies and eliminates records with minimal predictive value. This extraction process creates a reduced dataset that maintains prediction accuracy while significantly lowering computing resource requirements for the predictive analytics algorithm.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies preliminary data filtering and noise reduction actions before the predictive analytics algorithm processes the data. By pre-processing the data to remove irrelevant information and reduce dimensionality, the system prepares an optimized dataset that reduces the computational burden during the actual prediction process.

Inventive Principle:

Principle #10Preliminary action

3Productivity

If data filtering and noise reduction is applied, then prediction efficiency is improved, but data completeness may be compromised

Engineering Contradiction:

Improveprediction efficiencyVSAvoiddata completeness

Core Design Contradiction:

ProductivityVSLoss of information

Solution Approach 1:

The patent implements a feedback mechanism where the system continuously monitors prediction results and adjusts the data filtering process accordingly. The loop-back mechanism uses prediction accuracy metrics to refine the filtering criteria, ensuring that only truly non-impactful data is removed while preserving information that contributes to accurate predictions.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent employs dynamic data filtering that adapts to the specific characteristics of the input data and prediction task. The filtering criteria are not fixed but are adjusted based on the data distribution, prediction model requirements, and performance feedback, allowing the system to optimize the balance between data reduction and information preservation.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS11327938B2Method to improve prediction accuracy of business data with enhanced data filtering and data reduction mechanism

Publication Date: 2022.05.10 SAP SE

AI summary

According to some embodiments, a system associated with predictive analytics may include a data filter dictionary that stores a plurality of electronic data records based on prior input data. A data filtering and noise reduction module may receive input data and access the data filter dictionary. The data filtering and noise reduction module may then utilize data from the data filter dictionary and a latent semantic indexing data filter and noise reduction algorithm to remove information from the input data and create a subset of the input data. A predictive analytic algorithm platform may receive the subset of input data and use a predictive analytic algorithm to output a prediction result. An intelligent loop-back mechanism may then receive the subset of the input data and dynamically update the data filter dictionary based on an impact associated with the output prediction result.