Irregularly-Sampled Time Series Modeling with Spline Networks
This addresses the challenge of handling irregular and missing data in time series for applications like classification and forecasting, though it is incremental as it builds on spline-based imputation methods.
The paper tackled the problem of modeling irregularly-sampled time series with missing values by using splines as input to neural networks, applying transformations directly on interpolating functions instead of sampling points on a grid, resulting in competitive performance in accuracy and computational efficiency compared to existing methods.
Observations made in continuous time are often irregular and contain the missing values across different channels. One approach to handle the missing data is imputing it using splines, by fitting the piecewise polynomials to the observed values. We propose using the splines as an input to a neural network, in particular, applying the transformations on the interpolating function directly, instead of sampling the points on a grid. To do that, we design the layers that can operate on splines and which are analogous to their discrete counterparts. This allows us to represent the irregular sequence compactly and use this representation in the downstream tasks such as classification and forecasting. Our model offers competitive performance compared to the existing methods both in terms of the accuracy and computation efficiency.