Adaptive Network Predictor Switching HMM and GMM Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing time series prediction algorithms for network data become less accurate when the underlying structure of the data changes suddenly, leading to unpredictable patterns and correlations, making it difficult to forecast future conditions, especially with limited training data and resources.
Innovation Solution
An adaptive predictor system that switches between a primary predictor, such as a Hidden Markov Model (HMM) or Gaussian Mixture Model (GMM), and an alternative predictor, like a current data predictor, based on anomaly detection, allowing for more accurate predictions by adapting to changes in data patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single time series prediction algorithm is used, then the system is simple and easy to operate, but the prediction accuracy deteriorates when the underlying data structure changes suddenly
Solution Approach 1:
The system dynamically switches between a primary predictor and an alternative predictor based on anomaly detection. When anomalies are detected in the time series data, the system transitions from using the primary predictor to the alternative predictor, allowing it to adapt to sudden changes in data structure and maintain prediction accuracy without requiring a completely complex system architecture
Solution Approach 2:
An anomaly detection mechanism serves as an intermediary between the data input and the prediction process. This intermediary monitors the underlying structure of the time series data and triggers a switch between predictors when structural changes are detected, thereby protecting prediction accuracy without requiring the entire system to be overly complex
2Reliability
If more training data is collected to improve prediction accuracy, then the prediction model becomes more robust, but the time and resources required for training increase
Solution Approach 1:
The prediction system is segmented into multiple specialized predictors: a primary predictor trained on historical data for normal conditions, and an alternative predictor for anomalous conditions. This segmentation allows each predictor to be trained more efficiently on specific data types rather than requiring one massive predictor trained on all possible scenarios, reducing overall training time while maintaining accuracy
Solution Approach 2:
The system performs preliminary anomaly detection before making predictions. By detecting anomalies in the underlying structure of the time series data beforehand, the system can select the appropriate pre-trained predictor, avoiding the need to retrain models on-the-fly and reducing the time and computational resources required for adaptation
Data Source
AI summary
System and method embodiments are provided for adaptive anomaly detection based predictor for network data. In an embodiment, a computer-implemented method in a network component for predicting values of future network time series data includes receiving, with one or more receivers, network time series data; determining, with one or more processors, whether an anomaly is detected in the network time series data; generating, with the one or more processors, a prediction associated with the network data according to a primary predictor when no anomaly is detected in the network time series data; generating, with the one or more processors, the prediction associated with the network data according to an alternative predictor when an anomaly in the network time series data is detected; and sending, with one or more transmitters, the prediction to a network controller, wherein the network controller uses the prediction to adjust network parameters.