Joseph Cilli
Florida International University School of Computing and Information Sciences
Lecture Information:
- November 3, 2015
- 3:00 PM
- ECS: 349

Speaker Bio
Joseph Cilli received an Bachelor of Business Administration degree with
a minor in Computer Science from the University of Miami in 1994.
Joseph received a MasterÛªs degree in Computer Science 1995, also from
the University of Miami. After graduation, Joseph worked for a
non-profit organization for eight years as the Corporate Director of
Information Technology. For the past twelve years he has been employed
by Florida International UniversityÛªs Chaplin School of Hospitality &
Tourism Management, where he has served as the Director of Information
Technology and Director of Distance Education. While a part-time Ph.D.
student, Joseph has served as faculty member for the Chaplin School of
Hospitality and Tourism Management, teaching Hospitality Enterprise
Technology graduate students, as well as a teaching Ecommerce for
undergraduate students. He also earned an E-Learning Teaching
Certificate during this time from the University of Wisconsin-Stout.
Joseph has also designed professional development courses to assist in
faculty development in an online environment. He is an FIU Online 2.0
Founder and has developed various applications to assist with
instructional design. Joseph serves on the UniversityÛªs Faculty Senate
Online Committee, the Faculty Senate Technology Committee, and the ITAC.
JosephÛªs current research interests include Data Mining, Information
Retrieval, and Recommender Systems under the supervision of Dr. Tao Li.
Description
User-generated comments and reviews for hotels on the web are an
important information source for hoteliers and for travel planners. In
addition, understanding and reacting to these comments are important for
decision making and quality control. Online Travel Agents (OTA) have
gained in popularity over the last decade and have become the primary
source of consumer travel research and purchasing. Recent studies show
that 90% of consumers trust peer recommendations, while 70% trust online
comments left by strangers; whereas, only 14% trust advertisements.
This would suggest that user-generated comments left by strangers on
websites have significantly more value than company advertising
campaigns. It is determined that 97.7% of TripAdvisor.com users were
influenced by other traveler comments.
Recently, research in the areas of sentiment analysis and personalized
recommender systems have gained in popularity. As more people converge
on user-generated comment sites, the amount of information rapidly
increases to a point that is not manageable to neither hoteliers nor
consumers. Today, user-generated comment websites suffer from an
inability to satisfy the independent needs of individuals in a summative
manner. In general, sentiment analysis or opinion mining, refers to
text analysis with an objective of determining oneÛªs attitude with
regard to a specific topic or oneÛªs overall contextual polarity. A
recommender system is an information filtering technology whereby
information is refined by either content-based, collaborative, or a
hybrid of multiple systems. In such systems, consumers can quickly
evaluate a hotel based on the aggregation and summary of user-generated
comments.
This proposal will follow the flow of research that covers sentiment
analysis, recommender systems, and user profile creation. We will
investigate four aspects of sentiment analysis and recommender systems
which include (1) the sentiment analysis of user-generated comments in
the hotel industry; (2) the sentiment score of hotel specific attributes
and feature ranking (star rating); (3) the implicit creation of user
profiles based on previous comments; (4) a recommender system which
matches user profiles to hotel star ratings. We will also propose a
framework for a mobile-based, personalized hotel recommender application
based on (1) implicit and explicit user profiles; (2) hotel profiles
based on semantic generated star ratings; (3) a recommender system,
which matches user profiles and hotel profiles based on geographic position.