The Human, existed since millions of years and consequently, be inspired from the physiological phenomenon of the human body organs is something really interesting. This is the origin of the authors' new bio-inspired technique, called artificial haemostasis system (AHS), based on the haemostasis phenomenon that prevents and stops bleeding in case of external haemorrhage. Aiming at contributing to web searching they have applied their AHS to solve the problem of information retrieval following four steps: multilingual pre-processing (pre-haemostasis) to transform each text into a vector and ensure the service of multilingual search; The texts vectors pass through three filters: the primary information retrieval (primary haemostasis), the secondary information retrieval (secondary haemostasis) and the final information retrieval (fibrinolysis) using a selection step (plasminogen activation) to evaluate the relevance of each document to the query; the authors' experiments were performed on MEDLARS collection in order to show the benefit gained from using such approach compared to the classic one validated by a set of evaluation measures (recall, precision, FNR, FPR, f-measure, ROC, accuracy, Error, sensibility, and TCR); Finally, a result-mining step to see the results in graphical form with more realism, where the 3D cub method is largely preferred by the user than the cobweb method; The results of the system, are positive compared to the results provided by a conventional method and a set of bio-inspired techniques existed In literature (Simulating annealing (SA), Social worker bees (SWB), and Artificial social cockroaches (ASC)).
Evaluation Measures in Information Retrieval: A Theoretical Framework for the Unranked Retrieval Sets
