Estimating Top-k Destinations in Data Streams

摘要: One considers the problem of estimating most frequent values in a data stream. In many cases an approximate answer may be enough. A novel algorithm is presented to using mixed approach between counter-based techniques and sketch-based ones. The then used find destinations calls by individual customers telecommunications operators. use fast small footprint algorithms critical due huge number check answers are enough situations. that such detection needs performed for each customer kept up date at all times. This paper presents customer's behavior justify algorithms. Although this on well other contexts.