Machine Learning Model for Data-Analytic Visualization Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge lies in effectively visualizing large datasets, particularly in cybersecurity, where numerous factors and parameters must be considered to accurately identify relevant information, and a single choice in visualization can lead to either exposing important criteria or missing it, due to the complexity and overwhelming nature of the data.
Innovation Solution
A data analytics system captures input and output data from advanced user-generated visualizations, maps them into input/output data pairs, and trains a machine learning model to recreate effective visualizations, optionally applying thresholds to ensure only relevant visualizations are reported.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If traditional data processing methods are used to visualize large datasets, then the visualization process becomes manageable, but important criteria and patterns may be missed due to the overwhelming volume of data
Solution Approach 1:
The patent introduces an intermediary system comprising machine learning models and visualization generators that act as a mediator between the raw big data and the user. This intermediary automatically processes, analyzes, and generates visual representations of the data, reducing the complexity burden on users while ensuring that important patterns and criteria are not lost in the overwhelming data volume.
Solution Approach 2:
The system enables self-service by allowing the data to essentially visualize itself through automated machine learning models that independently identify patterns, generate visualizations, and present insights without requiring manual intervention. This reduces the complexity of manual data processing while ensuring comprehensive analysis of all data points.
2Measurement precision
If multiple visualization parameters and options are considered to accurately represent data, then the accuracy of information representation improves, but the complexity of selecting and configuring visualizations increases
Solution Approach 1:
The system automatically selects and configures the most appropriate visualization parameters and types by leveraging machine learning models that analyze the data characteristics and determine optimal representation methods. This eliminates the need for users to manually navigate complex visualization configuration options while maintaining high accuracy in information representation.
Solution Approach 2:
The system dynamically adjusts visualization parameters based on data characteristics, user preferences, and contextual factors. The machine learning models automatically modify parameters such as chart type, aggregation level, and display metrics to optimize the accuracy of information representation without requiring users to understand or configure these parameters manually.
3Ease of operation
If a single visualization choice is made to simplify the presentation, then the ease of understanding improves, but the risk of missing important criteria increases
Solution Approach 1:
The system segments the visualization process into multiple complementary views and perspectives, presenting several different visual representations of the same data simultaneously. Each visualization focuses on different aspects or criteria, allowing users to understand the data from multiple angles without being overwhelmed by a single complex view. This ensures that important criteria are not lost while maintaining ease of understanding through focused, simplified individual views.
4Reliability
If manual selection of visualization parameters is performed to ensure accuracy, then the relevance of information is improved, but the time and effort required increases
Solution Approach 1:
The system performs preliminary actions by pre-processing the data, pre-generating multiple visualization options, and pre-identifying relevant patterns and criteria before the user needs the visualizations. The machine learning models are trained in advance to recognize relevant information, so when visualizations are needed, they can be quickly generated with high relevance without requiring users to spend time on manual parameter selection and data analysis.
Data Source
AI summary
This document describes techniques and apparatuses for enhancing data-analytic visualizations of a data analytics system. A computing device captures input data and output data associated with a data-analytic visualization generated by an advanced user using the data analytics system. The input data and output data are mapped together for defining the data-analytic visualization generated. A machine learning model is trained relative to the mapped input data and output data for generating the data-analytic visualization. During a normal usage of the data analytics system by a user, the trained model generates data-analytic visualizations to suggest to the user responsive to input data from the user. An optional threshold is set and applied relative to the data-analytic visualizations generated. If a data-analytic visualization meets the threshold, the data analytics system reports the data-analytic visualization.