Text Field Classification for Dynamic IDs in Behavioral Biometrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing behavioral biometrics systems struggle to accurately identify and classify text fields with dynamically generated identifiers, leading to inefficiencies in keystroke dynamics analysis for authentication and fraud detection, especially when user interfaces change frequently.
Innovation Solution
A method and system for automatically generating unique identifiers for text fields by clustering metadata feature vectors into functional groups using statistical distance metrics, allowing for accurate classification and application of keystroke dynamics algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If random field identifiers are generated for each deployment or page load, then security of the webpage is improved, but the ability to determine field types and maintain behavioral biometrics profiles becomes difficult
Solution Approach 1:
The patent introduces an intermediary mapping system that connects random field identifiers to their functional types through metadata extraction and clustering. Instead of directly linking identifiers to field types, the system uses intermediate features (HTML tags, parent elements, contextual metadata) to bridge the gap, enabling automatic field type determination despite identifier randomization
Solution Approach 2:
The system changes the parameters used for field identification from relying on static identifier names to using dynamic contextual features and metadata. By extracting features such as HTML tag types, parent element relationships, and positional information, the system adapts to identifier changes while maintaining field type classification accuracy
2Measurement precision
If manual assignment of HTML IDs or CSS Tags is performed for each text field, then behavioral biometrics analysis accuracy is improved, but maintenance becomes infeasible when UI/app updates frequently
Solution Approach 1:
The system implements self-service by automatically extracting metadata and features from the user interface elements themselves. Instead of requiring manual annotation, the system queries the UI structure, extracts relevant features (tags, attributes, hierarchical relationships), and uses these to automatically classify fields and maintain behavioral profiles, making the system self-maintaining even with frequent UI updates
Solution Approach 2:
The system performs preliminary action by pre-extracting and storing metadata features from text fields during UI rendering or initialization. This preliminary extraction of contextual information (HTML tags, parent elements, positional data) enables later automatic field type determination without requiring manual intervention during maintenance or updates
3Adaptability or versatility
If field identifiers are dynamically generated, then adaptability to frequent UI updates is improved, but the ability to train and match behavioral biometrics sub-profiles becomes difficult
Solution Approach 1:
The patent segments the field identification process into independent feature extraction components (HTML tags, parent elements, positional information, metadata attributes). Each feature is extracted and stored separately, allowing the system to recombine these segments to identify fields regardless of identifier changes, thereby maintaining reliable behavioral profile matching through consistent feature-based recognition
Data Source
AI summary
Systems and methods are provided for automatically generating unique identifiers for text fields and determining the type of information of a specific text field in an online user journey. The method identifies a data type identifier for each text field by training on historical sessions from many user interactions to group text fields, even with randomly generated IDs, and based on the metadata associated with the text fields. The method can predict the data type for new data and can enable the automatic selection and application of a correct keystroke dynamics algorithm for authentication and fraud detection.