Institutional Repository
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/893756DF-A97E-46DA-9C2C-1D2024B108B7 | ||
| Year | 2018 | ||
| Type of Item | Diploma Work | ||
| License |
|
||
| Bibliographic Citation | Nektaria Mariolou, "A recommendation system for personalized reviews in Apache Spark", Diploma Work, School of Electrical and Computer Engineering, Technical University of Crete, Chania, Greece, 2018 https://doi.org/10.26233/heallink.tuc.78966 | ||
| Appears in Collections |
Personalized recommendation systems play an increasingly growing role in customer's decision making process. The various e-commerce sites differ in their objectives, functions and characteristics but their common primary goal is to identify efficiently user decision. The most frequent approach is the selection of the reviews with the highest percentage of helpfulness' votes by users who have read the reviews. Thus, they end up with a selection derived from a limited scope of criteria. In this work, we focus on retrieving a subset of reviews, using personalized criteria. In order to determine which set of reviews may correspond to individual users' preferences, we focus on the importance of product aspects for each user. Our system is built on Apache Spark enabling the processing of reviews, we evaluate it with a dataset from Amazon and the results are indexed in the distributed search engine, Elasticsearch
Copyright © DIAS 2013
