Format: ORAL

Valry Malcot

Département d'Ecologie, L'Institut Agro Rennes Angers, Angers, France

Learning plant names is usually considered as a complex and long process, which is even more complex when additional information is added (such as plant size, flowering period, leaf deciduousness, ). We will present the produce of our reflexions in the form of a database associated with a quiz generator, called TEPIK (for Test of Plant Identification and Knowledge). These refelxion have been done with tehe support of various professional groups and teachers. The produced tool, considered as a complement to teaching with living material, consist in a database fed by certified and free or licensed data. The main feeding sources are eflore from Tela Botanica and Floriscope from PlanteCit. Selection of the current taxa is based on wild and commercial frequency but any new taxa may be added if needed. For each taxa (more than 4000 are currently included) various characteristics and numerous pictures are present. Each picture is tagged in such way that the part of the plant presented is know by the system. Difficulty parameters are also implemented, whether they rely on the plant (i.e. some plant parts are more difficult to recognise), on questions (open questions are more difficult that single-choice questionnaire) or on information tested (knowing common name less difficult to learn than soil type). Teachers can generate sets of questions on a precise list of plants, with more or less detailed parameters in order to provide quiz to their students. In addition, thoughts on learning progression are currently developed with Bayesian Knowledge Tracing in order to follow and predict knowledge acquisition.