Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol.4, No.5, 2013
32
Research on Personalized Recommender System for Tourism
Information Service
Huang Yu
Yao Dan Luo Jing
Zhang Mu
*
Shenzhen Tourism College of Jinan University
6 QiaoCheng East Road, Overseas Chinese Town,Shenzhen 518053, Guangdong, China
* E-mail of the corresponding author: zhan[email protected]m
The research is supported by Science and Technology Planning Project of Guangdong province, China (Project
Number: 2012B031400008)
Abstract
Since the development in the 1990s, Recommender system has been widely applied in various fields. The
conflict between the expansion of tourism information and difficulty of tourists obtaining tourism information
allows Tourism Information Recommender System to have a practical significance. Based on the existing online
tourism information service and the mature recommendation algorithms, Personal Recommender System can be
used to solve present problems of the key recommendation algorithms. In the first place, this research presents an
overview of researches on this issue both at home and abroad, and analyzes the applications of main stream
recommendation algorithms. Secondly, a comparative study of domestic and international tourism information
service websites is conducted. Drawbacks in their applications are defined and advantages are adopted in the
settings of Recommender System. Finally, this research provides the framework of Recommender System, which
combines the design and test of algorithms and the existing tourism information recommendation websites. This
system allows customers to broaden experience of tourism information service and make tourism decisions more
accurately and rapidly.
Keywords: Tourism information service, Personalized recommendation, Intelligence recommendation module,
Apriori algorithm
1. Introduction
With the arrival of information age, tourism industry has expedited its pace of informatization. As a result,
tourism enterprises have grown increasingly dependent on information technology. Since tourists are getting
more and more sophisticated and rational, their needs tend to be diversified. Therefore, it is all the more
challenging to provide them with the requested tourism information. In the current domestic tourism market,
individualized travel is replacing traditional travel mode even though package tour is still the mainstream. To
participate in the individualized activities, tourists have to do a lot of inquiring and retrieving from the vast sea
of tourism information in order to get what they want, which hampers the development of personalized tourism
information service.
As a user-centered service mode, personalized tourism information service supplies tourism information and
services based on the user’s requirements, personality and travel habits. This system aims at providing valuable
information for the users’ reference when they make travel decisions. The personalized recommender system is
capable of catering for different needs, thus truly realizing the goal of service on demand. Personalized
recommendation of tourism information furnishes the development trend for tourism information.
2. Literature Review
The early recommender system focuses on content recommendation, and thus can’t do anything with such
information as music, image and video. To solve this problem, Konston
[1]
advances collaborative filtering
recommendation. It produces recommendations according to the similarity level of the users and other
parameters, so the recommendations are of higher value and timeliness.
In the research of travel recommender system, Schafer designs a system which simulates a travel agent who can
assist the user to get recommendation service online. David contrives an agent-based system named Intelligent
Travel Planning (ITP). It collects and processes travel information and recommends it to the user by dint of
intelligent agents with different functions. Making use of tourist-based textual response, Stanley
[2]
devises a
travel recommender system similar to a decision-support system. It presents information which may be of
interest to travel agencies and tourists. Recommender system develops with the application of artificial
intelligence. E-commerce platform such as Amazon is a case in point. Felfernig
[3]
is the first person who
brought to you by COREView metadata, citation and similar papers at core.ac.uk
provided by International Institute for Science, Technology and Education (IISTE): E-Journals
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol.4, No.5, 2013
33
advocates applying this technology to tourism, for instance, recommending destinations to mobile tourists.
Individuality, information filtering and recommendation are key technology in recommender system, among
which comparability is the most important concept. A mixed algorithm of comparability is put forward by
Zanker in allusion to information retrieval and CBR, and is put to use in E-tourism. Because traditional quotation
of tourism products is a comprehensive recommendation, Zanker
[4]
considers using constraint-based web
configuration and model-driven algorithm to obtain and maintain information, the core content of which is in
accordance with that of SOA. Srisuwan
[5]
sets and completes a personalized recommender system targeting
E-tourism, in which statistic technology based on Bayesian classification is utilized in providing
recommendation services for the users.
Currently domestic researches on personalized recommender system concentrate on the following aspects.Earlier
researches center on the comparative study of various recommendation algorithms with an emphasis on the
mainstream recommender system. By studying and analyzing various algorithms, the researchers put forward
suggestions for improvement and future research direction, such as Jianguo Liu
[6]
whose suggestions are based
on the characteristics and limitations of those commonly used systems, and Hailing Xu
[7]
who points out the
disadvantages and existing problems of different recommender systems through comparative analysis.Later
scholars focus more on improvement of recommendation algorithms. Mathematical methods are used in
improving algorithms. Zhi Zhao and Zhuonan Feng
[8]
analyze the existing problems with the traditional CF and
item-grade-based CF and put forward an optimized CF. Other researches are on the design of recommender
systems. Most of them build recommender systems from a macroscopic perspective and develop the systems’
functional blocks. You Lu and Li Yu
[9]
expound on the design process of intelligent recommender systems and
put forward some innovative ideas.
Another research area of interest with domestic scholars is personalized recommendation, which has been used
in E-commerce and web communities. The traditional algorithm of collaborative filtering can not reflect the
user’s interest change in a timely manner. In view of this problem, Chunxiao Xing and Fengrong Gao
[10]
from
Tsinghua University advance two measurements for improvement: time-based data weighting and
resource-similarity-based data weighting. These two weightings are introduced into the recommendation
production process of resource-based collaborative filtering algorithm. Experiments show that the improved
algorithm excels in recommendation accuracy. Guangwei Zhang
[11]
from Beijing University of Aeronautics and
Astronautics put forward a method which compares the knowledge similarities of the users, thus overcoming the
deficiency of the traditional method in measuring similarities. Focusing on this method, Zhang posts a new
collaborative filtering algorithm and proves its validity. Shouzhi Zhang and Yan Xu
[12]
from Fudan University
devise a personalized service system which realizes the dynamic drift of the user’s interest focus through statistic
analysis of the user’s behavior contrail.
3. Design of Personalized Recommender System for Tourism Information
3.1 Framework of Personalized Recommender System for Tourism Information
The goal of applying personalized recommender system to tourism information service is to provide tourists with
more efficient information searching experience and enhance individuation of tourism information.
In front of personalized recommender system, the user can input basic information, grade tourist attractions and
choose relevant types of travel through the interface. According to the inputs, the system produces specific
recommendations by using recommendation algorithm and then presents the results to the user in the form of
web page or e-mail (as shown in fig.1).
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol.4, No.5, 2013
34
Figure 1. Basic Framework of Recommendation System
(1) Presentation layer
The function of presentation layer is to enable the interaction between the user and the back stage and presents
the final results to the user’s inquiry. This layer can send the user’s requests to the server, gather data and display
information through websites. It acts as the interface for the user to access the system. By using general websites
as interfaces, it enables the user to visit websites whose operation system provides outward services and
demonstrates service information. Presentation layer mainly consists of GUI and the module of recommendation
page frame. GUI provides the user with imaging user interface and serves as the interface between the user and
the intelligent recommendation service module.
(2) Service logic layer
Service logic layer is made up of intelligent analysis module and intelligent recommendation service module.
Intelligent analysis module, made up of three modules of information collection, demand analysis and
recommendation set, is mainly responsible for analyzing the user’s requests whereas information collection
module mostly collects the user’s requests and personal information. Analytical module processes the collected
information. Then the module of recommendation set generates recommendations with the help of intelligent
recommendation service module.
Utilizing Apriori algorithm, intelligent recommendation service module picks out information in agreement with
the user’s needs from the user information database and tourism resource database and produces
recommendation set. It is a module capable of offering dynamic guidance for the user’s next step of operation
according to recommendation strategy and algorithm. Whether the strategy is reasonable has a direct impact on
the quality of the recommendation set.
(3) Database layer
Database layer consists of the user information database and tourism resource database. It is the storehouse for
operation data. Database layer stores and manages data through a nexus database.
Presentation layer, service logic layer and database layer are interrelated in realizing the function of providing
personalized recommendation service to the user.
3.2 Procedure of Personalized Recommendation
By posing requests to the website, the user enters the modules of information collection and demand analysis.
The inputted data will be processed by a combinational algorithm. The processed data then go to the data mining
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol.4, No.5, 2013
35
module, and finally the processed data will be stored in the user information database. In returning to the
recommendation set, relevant data will be extracted from tourism resource database. Through data mining and
computing, recommendation data will be produced, resulting in recommendation set which will get back to the
user via recommendation page frame (as shown in fig.2).
Figure 2. Work Procedure of Recommendation System
3.3 Main Features of the System
(1) Personalized Recommender System Has the Following Main Features:
Database is where the system stores its data. The data comes from different sources, including the user’s personal
information, purchasing and browsing records and tourism resource database. The user information database
stores users’ personal information inputted by them when they register, such as name, occupation, gender, hobby
as well as subject term, searching scope, frequency of the appearance of the key words, etc. Recommendations
will be made according to the users’ purchasing and browsing records to better meet their needs. On certain
occasions, the registration information can not exactly reflect the users’ traits because they skip some items
when inputting personal information. In this case, their purchasing and browsing records can help make out their
hobbies. Tourism resource database contains information about restaurants, hotels, transportation, sightseeing,
shopping and entertainments, ensuring satisfaction on the users’ part when they inquire information.
(2) The objective of this recommender system is to provide services which fulfill the users’ needs. Once a
request is sent, the system will automatically work to answer the call.
(3) The system can alter the recommendation service according to the changes of the registration information.
(4) The system realizes communication with the user in a timely manner. Since the database stores the user’ s
browsing record, it can obtain the user’s initial wants and likes. Without annoying the user, the system can
update relevant information according to the taxis of the user’s likes and dislikes regarding the services.
4. Design of the Function Module of Personalized Recommendation
4.1 Comparative Analysis of Recommendation Strategies
The choice of recommendation strategy determines the recommendation result, thus influencing the user’s
decision making. Table 1 shows the relationship between the existing recommendation strategies and the degree
of individuation.
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol.4, No.5, 2013
36
Table 1. Comparison of Recommended strategy
Recommended strategy
Degree of
Personalization
Degree of
Automation
Degree of
Persistent
Content-based retrieval Low Low Low
The top most active N-projects would be
Recommended
Low High Higher
The top best-selling N-projects which User
interested in would be Recommended
High High High
Content-based retrieval is a traditional searching technology. It is rather mature and widely used in libraries and
the searching system of various websites. The working principle is searching within the set scope for contents
which match with the target key words based on the subject and/or key words of the target content.
Currently the strategy adopted by many E-commerce websites is to recommend the first N items that are most
active. The results that are recommended to the user are the most active.
The third strategy combines the previous two. It strives to improve the demerit of inadequate individuation of
content-based retrieval strategy while keeping the merit of producing user-friendly recommendation results.
The core purpose of personalized recommender system is to satisfy the individual needs of different users.
Accordingly, the design of the system and the recommendation strategies should also reflect the mentioned
purpose. The recommendation method which integrates the user’s demands in the relevant tourist activities can
tie the user and the tourist attractions together, hence optimizing the recommendation strategy.
4.2 Design of Apriori Algorithm Module
4.2.1 Principle of this Algorithm
Apriori algorithm utilizes hierarchical sequential searching method to accomplish the mining of frequently
occurring information sets. K set is used to produce K+1 set.
Suppose I={i
1
,i
2
,…,i
m
} is the information aggregation of a tourism project, among which i
k
（k=1,2, …,m） is an
item. A transaction (T) is an item set, which is the sub-set of “I”. Every transaction is related to an exclusive
identifier TID. Normally two parameters are used to describe the attributes of Apriori algorithm.
(1) support
The support rate of Rule X=>Y in the database refers to the proportion between the number of transactions in the
trading set which contains both X and Y and the number of all transactions, marked as “support（X=>Y）”.
Support rate refers to the probability of transactions containing both X and Y.
(2) confidence
The confidence rate of Rule X=>Y in the transaction set refers to the proportion of transactions containing both
X and Y, namely, the probability of the occurance of B in transactions when A occurs, i.e. conditional
probability. It is often used to measure the validity of Apriori algorithm.
Support and confidence rates are two important concepts used to describe Apriori algorithm. The former reveals
the frequency of simultaneous appearance of X and Y. If the numerical value is small, it means the correlation
between X and Y is insignificant. The latter indicates whether Y will appear when X appears. Generally speaking,
only Apriori algorithm with both high support rate and high confidence rate would probably be what the user is
interested in.
If setting the transaction set D and the association rule X => Y of D, when c
≧
min_conf
∧
s the
≧
min_sup
and it means that X => Y is a strong association rules. In that, min_conf is minimum confidence and min_sup is
minimum support.
4.2.2 Overall Framework of Recommendation Module
Recommended module is the key components of the recommendation system (Figure 3 shows the design model
of this module). User can specify the min_sup, min_conf and respectively interact the Largest project set search
algorithm and Association rules algorithm. In the interaction, user can interpret and evaluate of the recommended
results.
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol.4, No.5, 2013
37
Figure 3. Overall Framework of Recommendation Module
This module includes two-step process. At first, it identifies the frequent item sets; therefore, it generates strong
association rules from frequent itemsets.
4.3 Application of Apriori Algorithm to Personalized Recommender System
For example, if we design a tourist information by assuming the min_sup=2, and min_conf = 40% (see Table
2).In this table, each item represents an attraction and each route consists of the various attractions. the
attractions characteristics could be extracted through the Recommendation module and then discover the
association rules between these characteristics by the Association rules algorithm.
Table 2. Tourism information
Travel Line Number Attractions
T1 I1,I2
T2 I1,I2,I4
T3 I2,I3
T4 I1,I3,I5
T5 I1,I2,I3,I4
T6 I2,I4
T7 I1,I2,I3
In this algorithm, the frequent 1 - itemsets L1={{I1},{I2},{I3},{I4},{I5}}, can be determined at first, then the
candidate set C2 by the minimum support. Finally, the frequent 3 - set L3={I1,I2,I3} could be found. By this
Iterative algorithm, all frequent item sets could be extracted (see Figure 4).
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol.4, No.5, 2013
38
Figure 4. Algorithm Process
After extract all frequent itemsets, the association rules could be generated by frequent itemsets {I1, I2, I3}
which including the non-frequent subset of {I1,I2},{I1,I3},{I2，I3},{I1},{I2} and {I3}.Following shows the
results according to the association rule generation method:
%50
4
2
}I,support{I
}I,I,support{I
conf,II∧I
21
321
321
====>
(Formula 1)
%67
3
2
}I,support{I
}I,I,support{I
conf,II∧I
21
321
231
====>
(Formula 2)
%67
3
2
}I,support{I
}I,I,support{I
conf,II∧I
21
321
132
====>
(Formula 3)
%33
6
2
}support{I
}I,I,support{I
conf,I∧II
2
321
312
====>
(Formula 4)
%40
5
2
}support{I
}I,I,support{I
conf,I∧II
1
321
321
====>
(Formula 5)
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol.4, No.5, 2013
39
%50
4
2
}support{I
}I,I,support{I
conf,I∧II
3
321
213
====>
(Formula 6)
According to the above results, because of the given min_conf = 40%, so only Formula 4 is not a strong
association rules. Thus the rest of the Formulas would be output through the personalized recommendation
system.
4.4 Model Testing and Experiment
4.4.1 Introduction of Weka
Weka (Waikato Environment for Knowledge Analysis) is the intelligent analysis system developed by Waikato
University. Weka provides a statistical interface, bringing together the most classic machine learning algorithms
and data processing tools (see Figure 5). Weka is an open platform, a collection of a large number of algorithms,
including classification, regression, clustering and association rules etc
[13]
.
Figure 5. Weka Interface
4.4.2 Apriori Algorithm Analysis
Open the contact-lenses.arff data under the data menu, then the Weka analyses and output the results thus
obtained the correlation between parameters.
(1) Open the contact-lenses data files in the Weka, the total of 24 records and 5 attribute values can be seen
(Figure 6).
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol.4, No.5, 2013
40
Figure 6. Contact-lenses Output Interface
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol.4, No.5, 2013
41
(2) Setting the Various parameters in the parameter interface ( see Figure 7).
Figure 7. Contact-lenses Parameters Interface
(3) To achieve complete data output, the output results and the results shown in Figure 8.
=== Run information ===
Scheme: weka.associations.Apriori -I -N 10 -T 0 -C 0.9 -D 0.05 -U 1.0 -M 0.5 -S -1.0 -c -1
Relation: contact-lenses
Instances: 24
Attributes: 5
age
spectacle-prescrip
astigmatism
tear-prod-rate
contact-lenses
=== Associator model (full training set) ===
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol.4, No.5, 2013
42
Figure 8. Result Analysis Output
Through above proceeding, we can verify that the associated algorithm for the analysis of the data processing to
calculate the confidence level through the Weka system output. Analysis of the output of the Weka system, we
can use a test model to analyses of tourism data sets so as to achieve the expected recommended result.
4.4.3 Weka System Experiments and Results Test
Embedding the Tourism data sets namely travel.arff into Weka, we use it to test whether it can be used in the
tourism information personalized intelligent recommendation module.
(1) Loading the travel.arff file in the Weka which consists of 19 properties(see Figure 9).
Apriori
=======
Minimum support: 0.5 (12 instances)
Minimum metric <confidence>: 0.9
Number of cycles performed: 10
Generated sets of large itemsets:
Size of set of large itemsets L(1): 7
Large Itemsets L(1):
spectacle-prescrip=myope 12
spectacle-prescrip=hypermetrope 12
astigmatism=no 12
astigmatism=yes 12
tear-prod-rate=reduced 12
tear-prod-rate=normal 12
contact-lenses=none 15
Size of set of large itemsets L(2): 1
Large Itemsets L(2):
tear-prod-rate=reduced contact-lenses=none 12
Best rules found:
1. tear-prod-rate=reduced 12 ==> contact-lenses=none 12 conf:(1)
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol.4, No.5, 2013
43
Figure 9. Travel Attribute Value Output Interface
(2) Running the data file, the output results can be shown in Figure 10 and Figure 11.
Figure 10. Travel Output Interface
=== Run information ===
Scheme: weka.associations.Apriori -N 10 -T 0 -C 0.6 -D 0.05 -U 1.0 -M 0.1 -S-1.0 -c -1
Relation: travel
Instances: 117
Attributes: 19
river
seaside
prairie
mountain
modern city
museum
revolutionary place
folk custom
tanker
religious sites
heritage
camping
theme park
polar region
rural scenery
climate places
creative places
art altar
total
=== Associator model (full training set) ===
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol.4, No.5, 2013
44
Figure 11. Apriori Algorithm Result
In Weka system, we can test out the minimum confidence of association rules between different attributes, so the
strongest association properties can be recommended by the minimum confidence.
Through the recommendation of tourist information services, if the recommended attractions in the same area,
the form of regional advantages, to provide better services for tourists, in order to attract more tourists and
improve the development of regional tourism. If recommended Scenic spots in different regions, through a
combination of tourist routes to improve the rationality of the design of the entire line.
5. Conclusion
The authors of this paper use Apriori algorithm to investigate the intelligent recommendation service in the
personalized recommender system for tourism information service. The research has accomplished the following
tasks:
(1) Based on the previous researches on tourism recommendation systems, the authors design a general frame for
the personalized recommender system in tourism information service. (2) After comparing and analyzing the
existing recommendation strategies, the authors integrate their superiorities and come up with personalized
recommendation strategies better adapted to tourism information service. (3) The authors mainly use Apriori
algorithm to complete the design of the intelligent recommendation module (4) Actual data are used to validate
the recommendation algorithm.
There exist deficiencies with this research. Firstly, there may be flaws with Apriori algorithm. The aggregation of
huge alternative choice items produced by the algorithm takes a lot of EMS memory, which goes against
efficient recommendation. Secondly, the personalized recommendation module designed by the authors is still
rather simplex in function and its link with the Internet is not powerful enough. Efforts will be made in the future
in perfecting the system structure so that it can recommend travel routes more efficiently with better results.
Acknowledge
This paper is supported by Science and Technology Planning Project of Guangdong province, China (Project
Apriori
=======
Minimum support: 0.1 (12 instances)
Minimum metric <confidence>: 0.6
Number of cycles performed: 18
Generated sets of large itemsets:
Size of set of large itemsets L(1): 12
Size of set of large itemsets L(2): 19
Size of set of large itemsets L(3): 2
Best rules found:
1. river=t seaside=t 24 ==> total=low 19 conf:(0.79)
2. polar region=t 23 ==> total=low 18 conf:(0.78)
3. seaside=t 68 ==> total=low 51 conf:(0.75)
4. seaside=t heritage=t 19 ==> total=low 14 conf:(0.74)
5. river=t 38 ==> total=low 27 conf:(0.71)
6. river=t total=low 27 ==> seaside=t 19 conf:(0.7)
7. theme park=t 19 ==> total=low 13 conf:(0.68)
8. heritage=t 40 ==> total=low 27 conf:(0.68)
9. mountain=t 18 ==> total=low 12 conf:(0.67)
10.total=low 78 ==> seaside=t 51 conf:(0.65)
Computer Engineering and Intelligent Systems www.iiste.org
ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online)
Vol.4, No.5, 2013
45
Number: 2012B031400008). This paper also supported by the Jinan University disciplinary research project
"Study on the Online Travel Information Search: a Perspective of Social Media".
References
Konston J,et al. GroupLens: Applying Collaborative Filtering to Usenet News. Communications of the
ACM,1997,40(3): 77-87.
Stanley Loh, Fabiana Lorenzi, Ramiro Saldana，Daniel Licthnow. A tourism recommender system based on
collaboration and text analysis. Information Technology & Tourism, 2003, 6(3):157-165.
Felfernig, A., Friedrich, G., Jannach, D., Zanker, M.: An Integrated Environment for the Development of
Knowledge-Based Recommender Applications, Intl. Journal of Electronic Commerce, Special issue on
Recommender Systems, 11(2):11-34.
Zanker M., Gordea S., Jessenitschnig M., Schnabl M.: A Hybrid Similarity Concept for Browsing
Semistructured Product Items. In: K. Bauknecht, B. Pröll, H. Werthner. (Hrsg.): Proceedings of 7th International
Conference on Electronic Commerce and Web Technologies - EC-Web '06. Berlin Heidelberg New York:
Springer 2006 (LNCS, 4082), pp. 21-30.
Srisuwan P, Srivihok A. Personalized trip information for e-tourism recommendation system based on Bayes
theorem[A]. //: Xu L, Tjoa A, Chaudhry S. Research and Practical Issues of Enterprise Information Systems Ⅱ:
IFIP International Federation for Information Processing. Boston: Springer, 2008. 1271-1275.
LIU Jianguo, Zhou Tao, Wang Binghong. Research Progress in Personalized recommendation system, Progress
in Natural Science, 2009,(1):1-15.(in Chinese)
XU Hailing, Wu Xiao, Li Xiaodong, Yan Baoping. Comparison study of internet recommendation system.
Journal of Software, 2009,(2):350-362. (in Chinese)
ZHAO Zhi; FENG Zhuo-nan. An adaptive algorithm of collaborative filtering recommender based on correlation
similarity.Journal of Changchun University of Technology(Natural Science Edition), 2006,27(4):354-358. (in
Chinese)
LV You; YU Li. Analysis and design of electronic commerce recommendation intelligence system.Agriculture
Network Information, 2006,(12):75-77. (in Chinese)
XING Chunxiao, GAO Fengrong, ZHAN Sinan, ZHOU Lizhu. A Collaborative Filtering Recommendation
Algorithm Incorporated with User Interest Change.Journal of Computer Research and Development,
2007,44(2):296-301. (in Chinese)
ZHANG Guang-Wei, LI De-Yi, LI Peng, KANG Jian-Chu, CHEN Gui-Sheng. A Collaborative Filtering
Recommendation Algorithm Based on Cloud Model, 2007,18(10): 2403-2411. (in Chinese)
ZHANG Shou-zhi; XU Yan. Design and Realization of a Personalization System.Mini-micro Systems,
2003,24(12):2155-2158. (in Chinese)
LIU Wen-Feng,QING Xiao-Xia. Study of Chameleon Clustering Algorithm and Implementation in
Weka.Computer Systems & Applications, 2010,19(12): 246-250. (in Chinese)
This academic article was published by The International Institute for Science,
Technology and Education (IISTE). The IISTE is a pioneer in the Open Access
Publishing service based in the U.S. and Europe. The aim of the institute is
Accelerating Global Knowledge Sharing.
More information about the publisher can be found in the IISTE’s homepage:
http://www.iiste.org
CALL FOR PAPERS
The IISTE is currently hosting more than 30 peer-reviewed academic journals and
collaborating with academic institutions around the world. There’s no deadline for
submission. Prospective authors of IISTE journals can find the submission
instruction on the following page: http://www.iiste.org/Journals/
The IISTE editorial team promises to the review and publish all the qualified
submissions in a fast manner. All the journals articles are available online to the
readers all over the world without financial, legal, or technical barriers other than
those inseparable from gaining access to the internet itself. Printed version of the
journals is also available upon request of readers and authors.
IISTE Knowledge Sharing Partners
EBSCO, Index Copernicus, Ulrich's Periodicals Directory, JournalTOCS, PKP Open
Archives Harvester, Bielefeld Academic Search Engine, Elektronische
Zeitschriftenbibliothek EZB, Open J-Gate, OCLC WorldCat, Universe Digtial
Library , NewJour, Google Scholar
