کد مقاله | کد نشریه | سال انتشار | مقاله انگلیسی | نسخه تمام متن |
---|---|---|---|---|
11012508 | 1798844 | 2019 | 21 صفحه PDF | دانلود رایگان |
عنوان انگلیسی مقاله ISI
Top-k frequent items and item frequency tracking over sliding windows of any size
دانلود مقاله + سفارش ترجمه
دانلود مقاله ISI انگلیسی
رایگان برای ایرانیان
موضوعات مرتبط
مهندسی و علوم پایه
مهندسی کامپیوتر
هوش مصنوعی
پیش نمایش صفحه اول مقاله
چکیده انگلیسی
Many big data applications today require querying highly dynamic and large-scale data streams to find the top-k most frequent items in the most recent window of a specified size at a specific time. This is a challenging problem. We propose a novel approach called Floating Top-k. Our algorithm does not need to explicitly maintain any item counts over time or deal with count updates upon item entry and expiration. Succinctly , we use only a small-size data structure to retrieve the top-k items dynamically in a window of any specified size within an upper bound. We prove that the space and time costs of Floating Top-k grow only logarithmically with the window size rather than the linear growth of previous methods. Our comprehensive experiments using three real-world datasets show that Floating Top-k not only provides accuracy guarantees but also has a memory footprint two to three orders of magnitude smaller and is one to two orders of magnitude faster than previous approaches. Hence, Floating Top-k is both effective and scalable, significantly outperforming competing methods. In addition, we devise a concise and efficient solution called Progressive Trend Model to address a related problem of tracking the frequency of selected items, improving upon previous methods by 20 to 30 times in model conciseness while maintaining the same accuracy and efficiency.
ناشر
Database: Elsevier - ScienceDirect (ساینس دایرکت)
Journal: Information Sciences - Volume 475, February 2019, Pages 100-120
Journal: Information Sciences - Volume 475, February 2019, Pages 100-120
نویسندگان
Chunyao Song, Xuanming Liu, Tingjian Ge, Yao Ge,