Encoded Session String Classification for Digital Component Distribution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for classifying user experiences with digital components rely heavily on click duration metrics, which can lead to erroneous classifications due to variations in user interactions and device conditions, failing to accurately assess the quality of experiences for activities like downloads, video views, and swipes.
Innovation Solution
A pattern-based evaluation system encodes online activities into symbols, generates encoded session strings, and classifies user sessions as positive or negative based on aggregated classifications, adjusting digital component distribution to improve user satisfaction without relying on click duration metrics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If click duration metrics are used to classify user experiences, then classification can be performed with simple data collection, but classification accuracy deteriorates due to variations in user interactions and device conditions
Solution Approach 1:
The patent segments user interaction evaluation into multiple independent dimensions (click duration, scroll depth, bounce rate, conversion actions) rather than relying on a single metric. Each dimension is encoded separately and combined to form a comprehensive classification, improving accuracy while maintaining manageable complexity through modular encoding functions.
Solution Approach 2:
The patent transforms the evaluation from using raw click duration values to using normalized encoded representations of multiple interaction parameters. By changing from a single parameter (click duration) to multiple parameters with different weights and encodings, the system achieves better classification accuracy while the encoding process standardizes the data to control complexity.
2Measurement precision
If comprehensive user interaction data is collected to improve classification accuracy, then user experience assessment improves, but data processing requirements and computational resources increase
Solution Approach 1:
The patent extracts only the most relevant interaction features from comprehensive user data through selective encoding functions. Rather than processing all possible interaction data, the system identifies and encodes key dimensions (click duration, scroll behavior, conversion actions) that contribute most to classification accuracy, reducing data volume while maintaining assessment quality.
Solution Approach 2:
The patent implements partial action by focusing on a subset of interaction dimensions that provide sufficient classification accuracy without processing excessive data. The encoding system selectively captures essential user experience indicators rather than attempting to process and analyze every possible interaction parameter, optimizing the balance between accuracy and data efficiency.
3Ease of operation
If traditional click duration-based classification is used, then implementation is simple and fast, but user experience quality assessment becomes erroneous for activities like downloads, video views, and swipes
Solution Approach 1:
The patent creates a universal encoding framework that handles multiple interaction types (clicks, downloads, video views, swipes) through a single multi-functional system. The encoding functions are designed to work across different interaction modalities, providing reliable quality assessment for diverse activities while maintaining implementation simplicity through consistent encoding patterns and a unified classification approach.
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for optimizing digital component transmission. A data structure stores session data for a user session. An encoder accesses the session data, encodes at least a portion of the online activities by representing different online activities with different symbols, and generates an encoded session string that includes multiple different symbols representing an order of occurrence of the different online activities. One or more servers classify the user session to a positive session classification or a negative session classification based on a classification score for the encoded session string, aggregates the classification of the user session with other classifications of other user sessions that have a same session string as the encoded session string, and adjusts distribution of digital components to client devices based on the aggregated classifications.