Adaptive Electronic Chat Splitting with Time-Gap Gaussian Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing e-discovery tools inefficiently process electronic chat data by treating entire chats as single documents, leading to inefficient manual review and inaccurate machine learning classification due to chats covering multiple topics, resulting in unnecessary processing of irrelevant content.
Innovation Solution
Adaptive splitting of electronic chats into conversations based on time gaps and Gaussian mixture models, allowing for precise indexing and classification of relevant content without requiring content analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If entire electronic chats are treated as single documents for e-discovery, then the document paradigm can be maintained and simplicity is preserved, but processing efficiency deteriorates and manual review becomes inefficient
Solution Approach 1:
The patent segments electronic chats into individual messages while maintaining the chat as a container. Each message is independently indexable and reviewable, yet all messages remain associated with the parent chat structure. This segmentation enables efficient manual review by allowing reviewers to focus on specific messages without abandoning the contextual framework of the entire chat.
2Device complexity
If entire electronic chats are indexed as single documents, then indexing simplicity is maintained, but search precision deteriorates when chats cover multiple topics
Solution Approach 1:
The patent segments chat content into individual messages for indexing purposes. Each message is independently analyzed and indexed with its own metadata, allowing search systems to precisely identify which specific messages contain relevant keywords. This maintains indexing simplicity at the message level while achieving high search precision by filtering individual messages rather than treating the entire chat as a monolithic unit.
3Device complexity
If machine learning classifiers process entire electronic chat documents, then processing is straightforward, but classification accuracy deteriorates when chats cover multiple topics
Solution Approach 1:
The patent segments electronic chats into individual messages that can be independently processed by machine learning classifiers. Each message is analyzed separately for classification, allowing the classifier to focus on the specific content of each message rather than being overwhelmed by the entire chat's multiple topics. This maintains processing straightforwardness at the message level while significantly improving classification accuracy by eliminating the dilution effect of irrelevant content.
4Loss of information
If reviewers must review entire electronic chat documents, then complete information is obtained, but time consumption increases unnecessarily
Solution Approach 1:
The patent segments chats into indexable messages with associated metadata, enabling selective review of only the messages that are relevant to the current task. Reviewers can efficiently navigate through segmented messages rather than reading entire chats from beginning to end, significantly reducing review time while maintaining information completeness through the structured organization and searchability of individual messages.
5Device complexity
If electronic chat data is stored as single documents, then storage simplicity is maintained, but downstream processing efficiency deteriorates
Solution Approach 1:
The patent segments electronic chat data into individual messages while preserving the parent-child relationship structure. Each message is stored as an independent, indexable unit with metadata linking it to the parent chat. This segmentation dramatically improves downstream processing efficiency for tasks like search, classification, and analysis, while storage remains relatively simple through the hierarchical organization of messages within chat containers.
Data Source
AI summary
Systems, methods, and computer program products for adaptively splitting electronic chats are provided. One embodiment includes receiving an electronic chat comprising a set of electronic chat messages, each of the electronic chat messages in the set of electronic chat messages having a timestamp; determining a set of time gaps between the electronic chat messages from the set of electronic chat messages; determining a set of models that model the set of time gaps, selecting an optimum model from the set of models; based on selecting the single Gaussian distribution as the optimum model, determining that the electronic chat comprises a single electronic chat, and storing the set of electronic chat messages as the single electronic chat.