Machine learning for transient discovery in Pan-STARRS1 difference imaging

TL;DR

This paper develops a pixel-based machine learning classifier to distinguish real astrophysical transients from false positives in Pan-STARRS1 difference images, aiming to automate transient identification for large sky surveys.

Contribution

It introduces a simple pixel-intensity based feature approach and compares multiple machine learning algorithms, achieving improved false positive rejection in transient detection.

Findings

Random forest achieved the best performance.

False positive rate of 1% yields about 10% missed detection.

Estimated missed detection rate is approximately 6%.

Abstract

Efficient identification and follow-up of astronomical transients is hindered by the need for humans to manually select promising candidates from data streams that contain many false positives. These artefacts arise in the difference images that are produced by most major ground-based time domain surveys with large format CCD cameras. This dependence on humans to reject bogus detections is unsustainable for next generation all-sky surveys and significant effort is now being invested to solve the problem computationally. In this paper we explore a simple machine learning approach to real-bogus classification by constructing a training set from the image data of ~32000 real astrophysical transients and bogus detections from the Pan-STARRS1 Medium Deep Survey. We derive our feature representation from the pixel intensity values of a 20x20 pixel stamp around the centre of the candidates. This differs from previous work in that it works directly on the pixels rather than catalogued domain knowledge for feature design or selection. Three machine learning algorithms are trained (artificial neural networks, support vector machines and random forests) and their performances are tested on a held-out subset of 25% of the training data. We find the best results from the random forest classifier and demonstrate that by accepting a false positive rate of 1%, the classifier initially suggests a missed detection rate of around 10%. However we also find that a combination of bright star variability, nuclear transients and uncertainty in human labelling means that our best estimate of the missed detection rate is approximately 6%.