GAME-Filter

Software capable of identifying microlensing events detected with Gaia telescope and deriving the properties of lensing objects that cause them

Collaborators: Taj Jankovič, Andreja Gomboc, Łukasz Wyrzykowsk , Uroš Kostić, Tanja Petrushevska, Mateusz Bronikowski, Maja Jabłonska, Zofia Kaczmarek, Martina Larma, Mile Karlica

Report a bug.

Table of Contents

1. About The Project
   • Prerequisites
   • Installation
2. Usage
• Basic Steps
• Running the code
3. Contact

About The Project

The Gaia mission's contribution to astrometry provides an unprecedented opportunity to observe microlensing events. While photometric detection has its merits, we focus on the identification of events through astrometric data. The aim is to leverage the astrometric precision of Gaia to develop GAME Filter, a software capable of identifying such microlensing events and to derive the properties of lensing objects that cause them.

(back to top)

Prerequisites

• Python 3
• astropy
• astromet
• jaxtromet
• scanning law

Installation

1. Download the Jupypter Notebook file

2. Install Python packages

(back to top)

Usage

Simulator (mock data; see Simulator_code)

• Generates Gaia mock observations for a range of single-source parameters (RA, DEC, pmra, pmdec, parallax) and saves them to .parquet files.
• Generates Gaia mock observations for a range of microlensing parameters (RA, DEC, pmra, pmdec, parallax, u0, thetaE, t0, tE, piEE, piEN) and saves them to .parquet files.
• Generates Gaia mock observations for a range of binary parameters (RA, DEC, pmra, pmdec, parallax, period, a, e, q, l, tperi, v_phi, v_omega, v_theta) and saves them to .parquet files.

GAME Filter (minimizer; see Minimizer_code)

GAME Filter (see Minimizer_code) is a software tool developed to identify microlensing events in the Gaia dataset and derive the properties of the lensing objects. The software reads xobs, xerr, Δxobs, tobs, and φobs from the Gaia data files. GAME Filter calculates xfit, the deviation along φobs at tobs, for specific single source and microlensing parameters. The software then minimizes a scalar parameter

MUWE = ( Σi=1N (xobs,i - xfit,i)2 / xerr,i2 )1/2,

which indicates the goodness of the microlensing fit. N corresponds to the number of observations for a specific event. The minimization process utilizes the Limited-memory Broyden-Fletcher-Goldfarb-Shanno algorithm (L-BFGS-B) to explore the parameter space and determine the optimal single source and microlensing parameters for individual events.

Following the minimization process, the minimizer might stop in an incorrect local minimum, failing to find the correct solution. Consequently, we establish criteria to determine when an event is recovered. These criteria are based on the value of MUWE after minimization MUWEmin, L2 optimality error Lopt, initial guesses, and the boundaries imposed on individual parameters. We consider an event as recovered if the following criteria are met:

• 0.9 < MUWEmin < 1.1.
• Lopt < 0.015.
• The values of πEE, πEN, and u0 differ from the initial guesses.
• The values of all parameters are within the imposed boundaries.

Contact

Taj Jankovič - taj.jankovic@ung.si

Project Link: [https://github.com/tajjankovic/GAME-Filter]

(back to top)