Word Prediction Using Non-Typing Digit Position
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current word prediction algorithms in electronic devices do not fully leverage positional data from non-typing digits on keyboards, limiting their ability to enhance text entry speed and accuracy.
Innovation Solution
Incorporating positional data from non-typing digits on keyboards to generate and weight word prediction candidates, adjusting weights for subsequent characters proximal to the detected position of the non-typing digit, and displaying these candidates to users for faster text input.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional word prediction algorithms are used without positional data from non-typing digits, then the system complexity remains low, but text entry speed and accuracy are limited
Solution Approach 1:
The system performs preliminary detection of non-typing digit positions on the keyboard and uses this information in advance to generate and weight word prediction candidates. By capturing the positional data before word completion, the system proactively prepares contextually relevant predictions, improving text entry speed without overwhelming complexity
Solution Approach 2:
The system continuously monitors the position of non-typing digits and uses this feedback to dynamically adjust the weighting of word prediction candidates. This real-time feedback mechanism refines predictions based on actual user typing patterns and finger positions, enhancing both accuracy and speed while maintaining manageable system complexity through iterative optimization
2Measurement precision
If traditional word prediction algorithms are used without positional data from non-typing digits, then the algorithm implementation remains simple, but text entry accuracy is limited
Solution Approach 1:
The system applies different weighting factors to word prediction candidates based on their proximity to the detected non-typing digit position. Candidates with subsequent characters corresponding to input members proximal to the non-typing digit receive higher weights, creating localized precision in predictions that matches actual user typing patterns and finger positions
Solution Approach 2:
The system adds a new dimension to traditional word prediction by incorporating spatial positional data from non-typing digits on the keyboard. This transforms the prediction from purely text-based to spatio-textual, using the physical location of fingers as an additional parameter to enhance prediction accuracy while maintaining algorithmic clarity
Data Source
AI summary
Methods and apparatuses are provided for improving word prediction in an electronic device. User input of one or more characters is received via a capacitive physical keyboard having a plurality of input members. Concurrently, the device determines the location of a non-typing digit, such as the user's finger or a stylus. Word prediction candidates are generated and weighted, and candidates that have subsequent characters associated with the input member proximal to the location of the non-typing digit are given more weight. The word prediction candidates are displayed, a second user input is then received comprising a selection of one of the candidates, and the device displays the selected word prediction candidate on the display.