FA-LAMP: FPGA-Accelerated Learned Approximate Matrix Profile for Time Series Similarity Prediction.
Amin KalantarZachary ZimmermanPhilip BriskPublished in: FCCM (2021)
Keyphrases
- financial time series
- prediction accuracy
- similarity measure
- real time image processing
- moving average
- chaotic time series
- similarity matrix
- prediction algorithm
- similarity function
- high speed
- distance measure
- prediction error
- light source
- distance computation
- box jenkins
- autoregressive
- longest common subsequence
- dynamic time warping
- prediction model
- singular value decomposition
- euclidean distance
- non stationary
- neural network
- real time
- mackey glass
- image processing
- forecasting accuracy
- knn
- low cost
- extreme values
- bit rate
- user profiles
- distance function
- exact solution
- symbolic representation
- distance metric