کد مقاله کد نشریه سال انتشار مقاله انگلیسی نسخه تمام متن
974174 1480137 2015 7 صفحه PDF دانلود رایگان
عنوان انگلیسی مقاله ISI
Predicting online ratings based on the opinion spreading process
ترجمه فارسی عنوان
پیش بینی رتبه بندی های آنلاین بر اساس روند گسترش افکار عمومی
کلمات کلیدی
سیستم توصیه شده، روند گسترش نظرات شبکه دو طرفه، فیلترینگ همکاری
موضوعات مرتبط
مهندسی و علوم پایه ریاضیات فیزیک ریاضی
چکیده انگلیسی


• We present a rating prediction method based on the user opinion spreading process.
• The user similarity is defined by measuring the amount of opinion a user transfers to another one.
• The algorithmic accuracy are improved greatly comparing with the item average method.

Predicting users’ online ratings is always a challenge issue and has drawn lots of attention. In this paper, we present a rating prediction method by combining the user opinion spreading process with the collaborative filtering algorithm, where user similarity is defined by measuring the amount of opinion a user transfers to another based on the primitive user-item rating matrix. The proposed method could produce a more precise rating prediction for each unrated user-item pair. In addition, we introduce a tunable parameter λλ to regulate the preferential diffusion relevant to the degree of both opinion sender and receiver. The numerical results for Movielens and Netflix data sets show that this algorithm has a better accuracy than the standard user-based collaborative filtering algorithm using Cosine and Pearson correlation without increasing computational complexity. By tuning λλ, our method could further boost the prediction accuracy when using Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE) as measurements. In the optimal cases, on Movielens and Netflix data sets, the corresponding algorithmic accuracy (MAE and RMSE) are improved 11.26% and 8.84%, 13.49% and 10.52% compared to the item average method, respectively.

ناشر
Database: Elsevier - ScienceDirect (ساینس دایرکت)
Journal: Physica A: Statistical Mechanics and its Applications - Volume 436, 15 October 2015, Pages 658–664
نویسندگان
, , , , ,