Adobe PDF (2.59 MB)
Title Details:
Evaluation Metrics
Authors: Symeonidis, Panagiotis
Description:
Abstract:
In chapter nine we will describe in detail the ways of evaluating the effectiveness (accuracy metrics, recall etc.) of recommendation algorithms. More specifically, we will describe the algorithms used to evaluate the algorithms. More specifically, in the introduction we will analyze the characteristics of off-line and on-line evaluation. Specifically for off-line evaluation, we will specify the separation of data into training and control. In addition, we will study the Mean Absolute Error and Root Mean Absolute Error metrics for predicting users' scores on the items. We will also present (using examples) the Confusion Matrix, Precision, Recall, Normalized Discounted Cumulative Gain and Receiver Operating Characteristic Curve (ROC curve) metrics, which specialize in evaluating the effectiveness of the proposed item list. Finally, we will also consider beyond accuracy metrics, such as the explainability metric and the novelty metric of the recommendation list items. With the former we evaluate whether a proposed item can be adequately explained to the user under consideration, while with the latter we evaluate whether the proposed item is novel with respect to the user's interests based on all of his/her previous interactions with the recommendation system.
Linguistic Editors: Sakellarios, Michalis
Technical Editors: Karatzidis, Dimitrios
Graphic Editors: Symeonidis, Panagiotis
Type: Chapter
Creation Date: 29-05-2023
Item Details:
License: Attribution - NonCommercial - ShareAlike 4.0 International (CC BY-NC-SA 4.0)
Handle http://hdl.handle.net/11419/9581
Bibliographic Reference: Symeonidis, P. (2023). Evaluation Metrics [Chapter]. In Symeonidis, P. 2023. Intelligent Recommender Systems [Postgraduate textbook]. Kallipos, Open Academic Editions. https://hdl.handle.net/11419/9581
Language: Greek
Is Part of: Intelligent Recommender Systems
Publication Origin: Kallipos, Open Academic Editions