Detecting false alarms in transit data from space: Rejection methods tested in Corot Blind Test 2

Abstract

Transit searches provide a large number of planet candidates. Before attempting follow-up observations, the best effort should be spent in classifying the light-curves, rejecting false alarms and selecting the most likely ones for real planets. A number of analysis tools has been developed with these objectives. Here, we apply such tools to 237 simulated multi-color light-curves from CoRoT Blind Test 2, which contain simulated planet transits and several configurations of impostors. Their comparison gives indications of the various tools' classification and false-alarm rejection capabilities. In order to arrive at the candidate identifications, we used an automated scheme of weighted punctuations assigned to the individual tests, which avoids that results from a single test dominate a candidate's classification.