Discover relaxed periodicity in temporal databases

The relaxed-periodicity pattern describes loose-cyclic behavior of objects while allowing uneven stretch or shrink on the time axis, limited noise, and inflation/deflation of attribute values. To discover relaxed-periodicity from temporal databases, we propose the concepts of attribute trend, trend inertia, peak-valley pattern, inertia algorithm with anti-noise ability as well as the peak-valley algorithm and show that the implementation prototype is efficient.