MLGBM Decision Trees for Optimal Call Timing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Communication service providers face low pickup rates and response fractions when contacting customers, as only a small percentage of calls are answered and responded to, necessitating an improvement in timing strategies for optimal contact.
Innovation Solution
The implementation of a modified Light Gradient Boosted Machine (MLGBM) system that builds decision trees using computational statistics to identify ideal times for contacting delinquent customers, incorporating attributes like credit scores and payment history to optimize call timing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional calling methods are used without optimization, then the calling system is simple to operate, but the pickup rate and response fraction are low
Solution Approach 1:
The system performs preliminary analysis of customer data, call history, and behavioral patterns before making calls to determine optimal calling times. This advance preparation enables higher pickup rates by contacting customers when they are most likely to answer, rather than using random or fixed scheduling approaches
Solution Approach 2:
The patent replaces traditional mechanical calling scheduling systems with an AI-based intelligent system that uses machine learning models to predict optimal calling times. This substitution of mechanical rules with intelligent algorithms enables dynamic optimization of call timing based on learned patterns from historical data
2Productivity
If AI-based optimization is implemented to predict ideal calling times, then the pickup rate increases significantly, but the system complexity increases
Solution Approach 1:
The AI system automatically trains and optimizes its own predictive models using historical call data and customer information without requiring manual intervention. The system self-adjusts to improving its accuracy over time by learning from past performance, reducing the need for complex manual configuration and maintenance
3Measurement precision
If more customer data attributes are analyzed to improve prediction accuracy, then the measurement precision of calling time prediction improves, but the computational complexity increases
Solution Approach 1:
The system extracts and focuses on the most predictive features from large amounts of customer data, such as call history patterns, time-of-day preferences, and demographic characteristics. By selecting only the most relevant attributes rather than processing all available data, the system achieves high prediction accuracy while managing computational complexity
Data Source
AI summary
A device comprises a processor. The processor is configured to: generate training vectors based on data related to communication with users; convert the training vectors into optimized vectors to be input into a machine learning unit; apply the machine learning unit to the optimized vectors to construct decision trees for determining probabilities of making a successful call during different time windows; generate a list pf calls and calling times based on the determined probabilities; and forward the list to an automatic dialer.