Automatic Construction and Natural-Language Description of Nonparametric Regression Models

Abstract

This paper presents the beginnings of an automatic statistician, focusing on regression problems. Our system explores an open-ended space of statistical models to discover a good explanation of a data set, and then produces a detailed report with figures and natural-language text. Our approach treats unknown regression functions nonparametrically using Gaussian processes, which has two important consequences. First, Gaussian processes can model functions in terms of high-level properties (e.g. smoothness, trends, periodicity, changepoints). Taken together with the compositional structure of our language of models this allows us to automatically describe functions in simple terms. Second, the use of flexible nonparametric models and a rich language for composing them in an open-ended manner also results in state-of-the-art extrapolation performance evaluated over 13 real time series data sets from various domains.

Cited by in corpus (26)

• One button machine for automating feature engineering in relational databases
• Differentiable Compositional Kernel Learning for Gaussian Processes
• Latent-space time evolution of non-intrusive reduced-order models using Gaussian process emulation
• StreetStyle: Exploring world-wide clothing styles from millions of photos
• Slow Momentum with Fast Reversion: A Trading Strategy Using Deep Learning and Changepoint Detection
• Tisane: Authoring Statistical Models via Formal Reasoning from Conceptual and Data Relationships
• Interpretable Distribution Features with Maximum Testing Power
• Uncertainty-Aware Calibration of a Hot-Wire Anemometer With Gaussian Process Regression
• Scalable Gaussian Processes for Characterizing Multidimensional Change Surfaces
• Big Learning with Bayesian Methods
• On the Semantic Interpretability of Artificial Intelligence Models
• client2vec: Towards Systematic Baselines for Banking Applications
• Time Series Structure Discovery via Probabilistic Program Synthesis
• Automatic Generation of Probabilistic Programming from Time Series Data
• Efficient Learning of Harmonic Priors for Pitch Detection in Polyphonic Music
• Probabilistic structure discovery in time series data
• How well does your sampler really work?
• The Automatic Statistician: A Relational Perspective
• Probabilistic Programming with Gaussian Process Memoization
• Gaussian Processes for Music Audio Modelling and Content Analysis
• Learning GPLVM with arbitrary kernels using the unscented transformation
• REMIX: Automated Exploration for Interactive Outlier Detection
• The Minecraft Kernel: Modelling correlated Gaussian Processes in the Fourier domain
• Heterogeneous Relational Kernel Learning
• Human-like Time Series Summaries via Trend Utility Estimation
• Inference of modes for linear stochastic processes