Miss-identification detection in citizen science platform for biodiversity monitoring using machine learning

摘要: Abstract In the recent years, several citizen science platforms for biodiversity monitoring have emerged. These represent a powerful tool collecting data researchers and increasing knowledge of participants. Typical are species names observed at given time place by numerous The use photos to document observations allows validation, in particular validation identification, key aspect needed quality control such databases. However, amount collected represents major challenge limited number co-opted experts dedicated validation. Therefore, detecting miss identifications can be very helpful focus expert workforce on dubious identifications. this paper, we test various machine learning approaches detect miss-identifications databases based features extracted form history validated observations. proposed model used automate process SPIPOLL platform.