Domain Name Suggestion Algorithm Using Session-Based Training Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for suggesting domain names to users lack effectiveness in personalizing recommendations based on user behavior, often mixing unrelated data from professional and personal activities, leading to suboptimal domain name suggestions.
Innovation Solution
A domain name registrar collects and logs user activities to create training data, using a learning algorithm that tokenizes and filters data to separate related domain name searches, suggestions, and registrations, ensuring accurate input-output pairs for training, and applies this data to an encoder-decoder model to suggest domain names.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If user activities are collected and stored in a training database for learning algorithm training, then the accuracy and personalization of domain name suggestions improve, but the risk of mixing unrelated data from different user activities increases
Solution Approach 1:
The patent segments user activities into distinct sessions using login sessions as boundaries. Each session's activities are processed separately to create training data pairs, ensuring that unrelated activities from different sessions are not mixed. This segmentation maintains data integrity while enabling comprehensive training data collection across multiple users and sessions.
Solution Approach 2:
The patent introduces tokenization as an intermediary processing step between raw user activities and training data creation. Domain names are converted into tokens, and related token sets are identified and grouped. This intermediary process filters and organizes data, preventing unrelated activities from being incorrectly associated while preserving relevant patterns for training.
2Reliability
If domain name searches are tokenized and filtered to separate related activities, then data quality for training improves, but the processing complexity and time required increases
Solution Approach 1:
The patent applies tokenization and filtering as preliminary actions during data collection and storage, rather than as post-processing steps. By converting domain names to tokens and identifying related activities upfront when creating training data pairs, the system ensures high data quality without requiring complex real-time processing during domain name suggestion generation.
3Adaptability or versatility
If training data is created from multiple user activities across different sessions, then the versatility and applicability of the learning algorithm improves, but the difficulty of ensuring data relevance and accuracy increases
Solution Approach 1:
The patent implements a dynamic approach to creating training data pairs by evaluating user activities in sequence within each session. The system adaptively determines which activities form valid input-output pairs based on temporal and contextual relationships, allowing the training data to dynamically reflect actual user behavior patterns while maintaining relevance through session-based boundaries.
Data Source
AI summary
Methods are taught for creating training data for a learning algorithm, training the learning algorithm with the training data and using the trained learning algorithm to suggest domain names to users. A domain name registrar may store activities of a user on a registrar website. Preferably, domain name searches, selected suggested domain names and domain names registered to the user are stored as the training data in a training database. The training data may be stored so that earlier activities act as inputs to the learning algorithm while later activities are the expected outputs of the learning algorithm. Once trained, the learning algorithm may receive activities of other users and suggest domain names to the other users based on their activities.