Struct PredictiveStreamingEngine 

pub struct PredictiveStreamingEngine<A: Float + Send + Sync> { /* private fields */ }

Predictive streaming engine for anticipating data patterns

Implementations

impl<A: Float + Send + Sync> PredictiveStreamingEngine<A>

pub fn new(config: &StreamingConfig) -> Result<Self>

pub fn predict_next( &mut self, current_data: &[StreamingDataPoint<A>], ) -> Result<Vec<StreamingDataPoint<A>>>

Train on the newly arrived samples, then forecast the continuation of the stream.

T8: every arriving sample is now actually used to fit the model (one exact recursive-least-squares update per observed coordinate) before any prediction is made, and predictions are only emitted once the model’s measured out-of-sample accuracy clears confidence_threshold. Previously the model was never trained and predict_next returned copies of the most recent observation, which carried no information about the future at all.

pub fn prediction_confidence(&self) -> A

The model’s current measured confidence in [0, 1].

pub fn prediction_error(&self) -> A

Measured out-of-sample mean absolute one-step prediction error.

pub fn normalized_prediction_error(&self) -> A

Measured prediction error relative to the scale of the observed data.

pub fn adaptation_rate(&self) -> A

The engine’s adaptation rate (how aggressively the fit chases a changing stream).