Speed Test Sentiment Prediction Using ML Feedback Gaps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing internet speed test systems lack sufficient customer feedback to accurately gauge user sentiment, limiting network operators' ability to optimize services and improve customer satisfaction.
Innovation Solution
A machine learning model is employed to predict user sentiment based on internet speed test data, using training data from multiple regions to account for varying infrastructure and user feedback patterns, and adjusting for outlier feedback to provide a more comprehensive understanding of user satisfaction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If customer feedback is collected through voluntary submissions, then implementation cost is low and system complexity is minimal, but the quantity of feedback is insufficient (only 2% of total speed tests)
Solution Approach 1:
The patent introduces a machine learning model as an intermediary between speed test data and sentiment analysis. The ML model processes speed test parameters (download speed, upload speed, latency, jitter) and predicts customer sentiment without requiring actual customer feedback submissions. This intermediary enables the system to generate sentiment insights from existing speed test data, dramatically increasing the effective quantity of sentiment data from 2% to potentially 100% of speed tests.
2Quantity of substance
If machine learning model is implemented to predict user sentiment, then quantity of sentiment data increases significantly, but device complexity and computational requirements increase
Solution Approach 1:
The patent applies partial action by implementing sentiment prediction only for speed tests where actual feedback is missing (the 98% gap). The system continues to collect actual feedback when submitted and uses ML prediction selectively to fill gaps, rather than replacing all feedback collection. This approach increases sentiment data quantity while limiting the increase in computational complexity to only the necessary extent.
3Ease of operation
If feedback collection relies on customer voluntary participation, then ease of operation is high, but the reliability of sentiment data is low due to insufficient sample size
Solution Approach 1:
The patent implements a feedback mechanism where the ML model's predicted sentiment scores are continuously refined based on actual customer feedback when it is submitted. The system uses actual feedback to train and validate the model, creating a closed-loop feedback system that improves reliability over time while maintaining the ease of voluntary participation. This feedback loop allows the system to leverage both the ease of voluntary collection and the reliability of actual sentiment data.
Data Source
AI summary
Methods and systems are described for predicting user (client or customer) sentiment associated with internet speed tests. In one embodiment, a system receives internet speed test data of an internet speed test performed on a data connection between a client device and a remote server. The system determines whether there is a user-provided sentiment score associated with the internet speed test data, and in response to determining that there isn't the user-provided sentiment score, produces, using the internet speed test data as input into a machine learning (ML) model, a predicted user-sentiment score as output of the ML model, wherein the predicted user-sentiment score relates to overall sentiment of a user of the client device with respect to the internet speed test data.