Sketch-Based Trend Querying for Time-Series Pattern Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing search tools lack the ability to accurately interpret and leverage the precise, quantified semantics of natural language inputs to identify and visualize temporal trends in time-series data, particularly in scenarios where users find it difficult to encapsulate their analytical intents through conventional textual inputs.
Innovation Solution
A system that enables users to search for trends in datasets by leveraging quantified semantics of language, incorporating a drawing interface to interpret sketches and natural language inputs, and using a semantic parser to detect analytical trend intents, with a search index that scores and ranks results based on label relevance and visual prominence.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional text-based search interfaces are used, then users can input queries, but the system cannot accurately capture nuanced analytical intents and temporal trend patterns
Solution Approach 1:
The patent introduces sketch-based input as a new dimension alongside text-based search. Users can draw temporal patterns directly on a canvas, transforming abstract analytical intents into concrete visual representations. This dimensional addition enables precise capture of trend patterns (increases, decreases, seasonality) without relying on limited textual descriptors.
Solution Approach 2:
The system employs an intermediary processing layer that translates sketch drawings into structured query parameters. The sketch parser converts visual patterns into machine-interpretable representations, bridging the gap between user intuition and system execution. This intermediary transformation enables accurate trend detection while maintaining ease of operation through intuitive drawing interfaces.
2Adaptability or versatility
If sketch-based input is introduced, then nuanced analytical intents can be captured, but the system complexity increases
Solution Approach 1:
The patent segments the sketch processing into distinct functional modules: sketch parsing, pattern recognition, query generation, and result retrieval. Each module handles a specific aspect of the complex task independently. The sketch parser handles raw input, pattern recognition identifies trends, query generation formats requests, and retrieval executes searches. This segmentation reduces overall system complexity by making each component manageable and testable.
Solution Approach 2:
The sketch-based interface serves multiple functions: it can represent various trend patterns (increases, decreases, plateaus, seasonality), work with different data types, and accommodate diverse user skill levels. The universal sketch parser handles all these cases through a single unified processing pipeline, avoiding the need for separate specialized systems for each query type while maintaining high adaptability.
3Productivity
If conventional search tools are used, then basic data retrieval is possible, but temporal trend analysis and pattern identification are limited
Solution Approach 1:
The patent replaces conventional mechanical text-matching search with a visual pattern recognition system. Instead of relying on keyword matching and textual metadata, the system uses sketch-based visual input to directly capture temporal patterns. This substitution enables both efficient data exploration through intuitive drawing and reliable trend identification through direct visual pattern matching, overcoming the limitations of text-based approaches.
Data Source
AI summary
A computing device receives, via a user interface, a drawing input directed to a dataset of time series data. The computing device converts the drawing input into a set of search terms and executes a query against a search index for the dataset using the set of search terms to retrieve a plurality of labeled trend events. Each of the labeled trend events corresponds to a respective portion of a respective line chart of a set of line charts representing the time series data and has a respective chart identifier. The computing device generates a first subset of line charts according to the retrieved plurality of labeled trend events. The computing device displays, on the user interface, one or more line charts of the first subset of line charts.