An inverted index method for mass spectra k-nearest neighbor que...

Finding k-nearest neighbors (k-nn) in metric-space is frequently used in modern biological applications due to its general applicability. Processing such queries with general purpose methods usually requires more time and space than domain-specific methods. This paper presents an inverted index method which exploits the sparsity of mass spectra binary format data and compares it with an existing metric-space method. This metric-space method acts as a coarse filter and can be followed by any fine