Boosted Embeddings for Time Series Forecasting

TL;DR

This paper introduces DeepGB, a novel time series forecasting model that combines gradient boosting with deep neural network embeddings, outperforming existing models on real-world datasets.

Contribution

The paper presents a new embedding architecture within a gradient boosting framework for deep learning models in time series forecasting, enhancing performance.

Findings

DeepGB outperforms state-of-the-art models on sensor data.

The new embedding architecture improves forecasting accuracy.

Gradient boosting with DNNs effectively captures complex time series patterns.

Abstract

Time series forecasting is a fundamental task emerging from diverse data-driven applications. Many advanced autoregressive methods such as ARIMA were used to develop forecasting models. Recently, deep learning based methods such as DeepAr, NeuralProphet, Seq2Seq have been explored for time series forecasting problem. In this paper, we propose a novel time series forecast model, DeepGB. We formulate and implement a variant of Gradient boosting wherein the weak learners are DNNs whose weights are incrementally found in a greedy manner over iterations. In particular, we develop a new embedding architecture that improves the performance of many deep learning models on time series using Gradient boosting variant. We demonstrate that our model outperforms existing comparable state-of-the-art models using real-world sensor data and public dataset.