sustainability

Article
Mobile e-Commerce Recommendation System Based
on Multi-Source Information Fusion for
Sustainable e-Business
Yan Guo 1 , Chengxin Yin 2, *, Mingfu Li 2 , Xiaoting Ren 3 and Ping Liu 4
1 College of Management Science, Chengdu University of Technology, Chengdu 610059, China;
[email protected]
2 College of Architecture Engineering, Chengdu Aeronautic Polytechnic, Chengdu 610100, China;
[email protected] or [email protected]
3 College of Public Management, China University of Mining and Technology, Xuzhou 221116, China;
[email protected]
4 Department of Economic and Management, Sichuan Technology Business College, Chengdu 611830, China;
[email protected]
* Correspondence: [email protected] or [email protected]

Received: 30 November 2017; Accepted: 7 January 2018; Published: 9 January 2018

Abstract: A lack of in-depth excavation of user and resources information has become the main
bottleneck restricting the predictive analytics of recommendation systems in mobile commerce.
This article provides a method which makes use of multi-source information to analyze consumers’
requirements for e-commerce recommendation systems. Combined with the characteristics of mobile
e-commerce, this method employs an improved radial basis function (RBF) network in order to
determine the weights of recommendations, and an improved Dempster–Shafer theory to fuse the
multi-source information. Power-spectrum estimation is then used to handle the fusion results and
allow decision-making. The experimental results illustrate that the traditional method is inferior to
the proposed approach in terms of recommendation accuracy, simplicity, coverage rate and recall rate.
These achievements can further improve recommendation systems, and promote the sustainable
development of e-business.

Keywords: mobile e-commerce; information system; predictive analytics; information technology;

data-mining; decision-making

1. Introduction
Mobile e-commerce is derived from the broader concept of e-commerce and refers to business
activities conducted by mobile terminals, such as mobile phones and tablet computers [1,2]. With changes
in user consumption habits, mobile e-commerce has become a trend. However, the continuous generation
of a large quantity of data is not only inconvenient for consumers when seraching for meaningful products,
but also means some of the products are rarely purchased. Meanwhile, the limited mobile terminal visual
interface greatly affects the consumer’s shopping experience and satisfaction. These factors may alienate
customers and impede the sustainable development of e-business. As an information technology to resolve
the above problems [3,4], the recommendation system is widely applied by e-commerce practitioners
and has become an important research topic in information science and decision-support systems [5,6].
Currently, the research on recommendation systems generally includes content-based filtering (CBF) [7],
collaborative filtering (CF) [8,9], and other data-mining techniques [10], such as decision trees [11,12],
association rules [13], and the semantic approach [14]. Although recommendation systems that could
help users find the information they need more quickly have received extensive attention [15], research

Sustainability 2018, 10, 147; doi:10.3390/su10010147 www.mdpi.com/journal/sustainability

Sustainability 2018, 10, 147 2 of 13

regarding mobile e-commerce recommendation systems continues to lag in two aspects that motivate
our study.
First, the e-commerce recommendation system is not deep enough to mine consumers’ online
behavior in multi-source mining. Recommendation systems only focus on information about products
and consumers’ shopping behavior on the shopping platform. The systems do not combine other
information in the mobile phone: pictures, music, video, applications etc. Therefore, the accuracy of
the recommendation is limited. It is like the story of the blind men and the elephant: a blind man
feels an elephant, only touching some part of it and then concluding what the elephant is like. That is,
he draws a conclusion from incomplete data. Mobile-phone users produce a lot of information every
day, not just shopping information on the shopping platform but also other platforms such as social
media sites [16,17]. We can obtain information from various sources in order to understand consumer
needs, so as to recommend products. Let us consider a scenario below. You see a trendsetter wearing
a pair of good shoes in the street and take a photo. Then, the mobile e-commerce recommendation
system recommends a shop for you, and you can also buy a pair of shoes for yourself.
Second, the mobile e-commerce recommendation system should integrate user location information.
Localization is one of the characteristics of mobile e-commerce. There is a close relationship between
the consumer’s location and their shopping behavior. Traditional e-commerce makes it difficult to
collect consumer information anytime and anywhere. A mobile terminal can remove the shackles of the
computer. Positioning technology allows the recommendation system to facilitate consumers when it
comes to mobile business, transactions and services. In order to let the recommendation system play an
enhanced role, a multi-source information-fusion technique should be imported into the integration of
location and historical behavior information.
This paper proposes a recommendation method using multi-source information-fusion technology
based on consumer location information and online behavior in the mobile terminal. First, this research
designs a flow chart of a mobile e-commerce recommendation system. Then, the paper extracts evidence
and uses the improved radial basis function (RBF) neural network to set weights. Based on that,
the article uses the improved Dempster–Shafer (D–S) evidence theory to fuse the information and
the power-spectrum estimation in order to recommend products. Finally, a Taobao online dress shop is
used as the case study to testify the validity of the proposed mobile e-commerce recommendation system.
The paper provides a detailed operation method and experimental results. The rest of the paper is as
follows: Section 2 reviews the literature on recommendation systems, multi-source information-fusion
techniques, the RBF neural network, and D–S evidence theory. A method to address the problem is
presented in Section 3. Section 4 uses a case study to prove the performance of the proposed method.
Conclusions and discussion are provided in Section 5.

2. Literature Review
The literature review focuses on the following aspects: recommendation systems, multi-source
information-fusion techniques, the RBF neural network, and D–S evidence theory.

2.1. Recommendation Systems

Recommendation systems are tools to aid decision-making that analyze a customer’s previous
online behavior and recommend products to meet their preferences [18–20]. E-commerce retail
websites have widely used them to provide consumers with personalized marketing services in
recent years [21]. Since Goldberg and his colleagues [22] first proposed recommendation systems,
other researchers have proposed a wide range of recommendation systems and related technologies,
including CBF, CF, and other data-mining techniques [23]. For example, Pazzani and Billsus introduced
the strategy architecture of the content-based method [24]. Unlike earlier research that focused on
technical factors [25], other studies turned their attention to the impact of recommendation systems on
a customer’s buying decision [8,26–28]. Although there are many papers related to recommendation
models, strategies [29], or procedures, few works have been conducted for a mobile e-commerce
Sustainability 2018, 10, 147 3 of 13

recommendation system, which has specific characteristics. In existing research, scholars have collected
the recommendation information solely from the shopping platform instead of multiple information
channels in the mobile terminal.

2.2. Multi-Source Information-Fusion Technique

Since information fusion was formally proposed in the 1990s, multi-level processing in the process
of information fusion has become the consensus among scholars of the information-fusion model.
Yager [30], Gregor [31] and many other scholars [32–34] have researched in depth the multi-source
information-fusion framework, information classification, automatic reasoning, heterogeneous data
processing, cloud computing and the peer-to-peer (P2P) network trust model of information fusion.
Most of the information-fusion models are based on the joint directors of laboratories (JDL) model
established by the United States Department of Defense, which realizes the requirement of multi-source
information fusion from four different levels of processing. As multi-source information-fusion
technique research has developed, it has been used in pattern recognition, data mining, knowledge
discovery and so on. However, there has been less research of multi-source information fusion based
on location in mobile e-commerce recommendation systems.

2.3. Radial Basis Function (RBF) Neural Network and Dempster–Shafer (D–S) Evidence Theory
A RBF neural network is a radial basis function neural network that is an efficient feedforward
neural network (FFNN) and can approach any non-linear function. It can handle the difficulty of
analyzing the regularity of a system, has a good ability to generalize and a fast convergence rate.
Researchers have succeeded in bringing RBF neural networks to the following fields: non-linear function
approximation, pattern recognition, time-series analysis, image processing, data classification, system
modeling, information processing, and control and fault diagnosis. For example, Yun et al. established a
short-term load-forecasting model using a RBF neural network [35]. However, the results of classification
based on a RBF neural network often appear negative [36,37], which does not support evidence well.
Yingwei et al. analyzed the minimal resource allocation network (M-RAN) learning algorithm which
has fewer hidden neurons but better classification accuracy [38]. Huang et al. proposed a new sequential
learning algorithm for RBF networks for function approximation [39]. This paper establishes an improved
RBF neural network model, and uses it to determine the weight of the evidence.
Evidence theory is a kind of inexact inference theory. After Dempster first put forward evidence
theory in 1967, it was further developed by his student Shafer [36], also known as the Dempster–Shafer
(D-S) evidence theory. The main purpose of evidence theory is to synthesize uncertain data from
many sources. Evidence theory uses the formula of evidence synthesis to compose the credibility of
different information sources so as to obtain decision-making results [40]. Skowron and Grzymala [41]
proposed a method of evidence acquisition based on a decision table. Yang et al. [42] studied the
evidence acquisition method for a large decision table. Deng et al. [43] studied the synthesis method of
evidence theory, and designed the expert weight measurement method based on the distance required
to effectively synthesize conflicting evidence. However, D–S evidence theory has fairness, one-vote
veto, robustness and Zadeh paradox problems [44,45] in its application. Scholars have improved the
method mainly through modifying the data source and the composition rules [46,47]. This paper
adopts evidence-body modification based on credibility and synthesis rules that are based on local
conflict allocation in order to fuse evidence.

3. Modeling

3.1. The Solution Framework

The detailed steps of the product recommendation method based on multi-source information
fusion are stated in the following four sub-sections, including data sources, recommendation evidence
weight, fusion decision, and result. The flow chart is shown in Figure 1.
Sustainability 2018, 10, 147 4 of 13
Sustainability 2018, 10, 147 4 of 13

Figure 1.
Figure 1. Flow chart
chart of
of aa mobile
mobile e-commerce
e-commerce recommendation
recommendation system.
system.

3.2. Data
3.2. Data Sources
Sources
The mobile
The mobile e-commerce
e-commerce recommendation
recommendation system system needs
needs accurate,
accurate, unified
unified consumer
consumer information
information
(features and needs). However, current consumer information management is usually in thethe
(features and needs). However, current consumer information management is usually in state
state of
of an
an “isolated
“isolated island”;
island”; that that
is, theis,complete
the complete information
information of a mobile
of a mobile user isuser is separated
separated and respectively
and respectively stored
stored in the mobile terminal, which lacks a unified representation and interface.
in the mobile terminal, which lacks a unified representation and interface. To solve the above problem, To solve the above
problem,
the proposed therecommendation
proposed recommendation system uses
system uses Micro-formats Micro-formats
for user for user information
information representation, storage,
representation, storage, integration and management. Although foreign
integration and management. Although foreign business has widely used Micro-formats to integrate business has widely used
Micro-formats to integrate information of a blog or micro-blog, and
information of a blog or micro-blog, and achieved success in YouTube, Flickr, MapQuest and MySpace achieved success in YouTube,
Flickr, MapQuest
applications, and MySpace
the anonymity of theapplications, the anonymity
external network leads to roughof theandexternal
fragmentednetwork leads
user to rough
information,
and fragmented user information, making it difficult to give full play
making it difficult to give full play to the effectiveness of Micro-formats. Mobile-phone applications to the effectiveness of
Micro-formats. Mobile-phone applications have widely implemented
have widely implemented a real-name registration system and unification of the mobile phone a real-name registration
system and
number. Theunification
recommendation of the mobile
systemphone number.
can obtain moreThe recommendation
accurate and detailedsystem consumer can information,
obtain more
accurate and detailed consumer
so Micro-formats are very suitable for it. information, so Micro-formats are very suitable for it.
Asmentioned
As mentioned above,
above, somesomeof the of the consumer
consumer information information
(such as gender,(suchage,aslevel
gender, age, leveletc.)
of consumption, of
consumption,
required by the etc.) required by system
recommendation the recommendation
can be obtained systemfrom thecan be obtained
shopping platform from the shopping
through a timing
update. However, more user information is still being explored; for example, the immediateexplored;
platform through a timing update. However, more user information is still being needs of the for
example,the
product, thepotential
immediateneeds, needs
recentof attention,
the product,
etc. the
Thispotential needs,that
article assumes recent attention, etc.systems
recommendation This article
can
assumes that recommendation systems can collect the information they need.
collect the information they need. Location information, voice, picture, video and so on can be collected in Location information,
voice,
the picture,
setting of thevideo
mobile and so onFor
phone. canother
be collected in the setting
app information, of thethat
we assume mobile phone. sharing
information For other canapp
be
information, we assume that information sharing can be achieved in order
achieved in order to provide the consumer with timely recommendation services. For example, the system to provide the consumer
with timely
collects recommendation
a user’s needs and concerns services. For example,
in a social the system
app. The system collects
can extract a user’scharacteristics
personal needs and concerns
and the
in a social app. The system can extract personal characteristics
recent focus of the consumer by means of “social platform nickname–mobile phone number”. and the recent focus of the consumer
Furthermore,
by means
the system of “social the
excavates platform nickname
consumer’s –mobile
historical phoneinformation
purchase number”. Furthermore,
on the shopping theplatform
system excavates
database
the consumer’s historical purchase information on the shopping
through the “mobile phone number—shopping platform ID”. Finally, the two kinds of information platform database through are the
“mobileinto
injected phone number—shopping
the user’s information as platform
evidence. ID”. Finally,includes
The process the two several
kinds of information
steps, as shownare injected
in Figure 2.
into the user’s information as evidence. The process includes several steps, as shown in Figure 2.
Sustainability 2018, 10, 147 5 of 13
Sustainability 2018, 10, 147 5 of 13

User 2.
Figure 2.Figure information acquisition
User information and integration
acquisition of Mobile
and integration e-commerce
of Mobile recommendation
e-commerce recommendationsystem.
system.
3.3. Recommendation Evidence Weight
3.3. Recommendation Evidence Weight
An improved
An improvedRBF neural network
RBF neural is isdesigned
network designed based
based on onthethe
JavaJava neural
neural network
network packetpacket
Joone toJoone to
solve thesolve the shortcomings of the RBF neural network mentioned in Section 2.3. The input layer oflayer
shortcomings of the RBF neural network mentioned in Section 2.3. The input the of the
network is still the linear input LinearLayer; the hidden layer is the RbfGaussianLayer; the output
network is still the linear input LinearLayer; the hidden layer is the RbfGaussianLayer; the output
layer is layer is the SoftmaxLayer;
the SoftmaxLayer; andand thenetwork
the network layer
layerand
andthethe
layer-weight
layer-weighttraining functionfunction
training are based are
on based
the elastic feedback algorithm Resilient Backpropagation (RPROP) [48,49]. The RbfGaussianLayer
on the elastic feedback algorithm Resilient Backpropagation (RPROP) [48,49]. The RbfGaussianLayer
transfer function is a Gauss-type function seen in the following form (1):
transfer function is a Gauss-type function seen in the following form (1):
2
x − cj
y j = exp(−k x −2 cj k
)2 (1)
y j = exp(− σ 2j ( x) ) (1)
σj ( x )
The central position c j and variance σ2j (x) are set by calculating the mean value and variance
Theofcentral
the input. The number
position of hidden
c j and variance σj2 (inx )the
layers network
are set byneeds to be optimized.
calculating the mean Thevalue
outputand
layer,
variance
SoftmaxLayer, adopts the transfer function shown in Formula (2), where c is
of the input. The number of hidden layers in the network needs to be optimized. The output layer, the number of output

SoftmaxLayer, adopts they transfer

classifications, and i is the neuron output.
function shown in Formula (2), where c is the number of output
classifications, and yi is the neuron output. exp( xi )
yi = c

(2)
exp(
exp ( xix)i )
yi = ic=1 (2)
∑ exp( xi )
In decision analysis, different sources of information have a different importance, which in turn
i =1
play different roles in our decision making. Therefore, the weights of different information sources
In decision analysis,
are different different sources
in the recommendation of information
system. The trainedhave
neurala different
network modelimportance,
is used which
for the in turn
play different
mobile roles in our recommendation
e-commerce decision making. Therefore,
system. the weights
The output is Y = {yof
net net different information
net net
1 , y2 ,..., ysum}
sources are
( sum is the total
differentnumber
in theofrecommendation
information sources). system. The trained
In the article, neuralasnetwork
Ynet is defined model is used forbased
the mobile
the recommendation weight
e-commerce recommendation system. The output is Y net = y1net , y2net , . . . , ynet

(sum isnetthe total
on an improved RBF network and used for information fusing as evidence. The sum
element y i
number of information sources). In the article, Ynet is defined as the recommendation weight based
reflects the degree of relative support of the recommendation information between information
on an improved RBF network and used for information fusing as evidence. The element yinet reflects
sources, and satisfies sum y net = 1 .
the degree of relative support 
i =1
i
of the recommendation information between information sources,
sum
and satisfies ∑ yinet = 1.
i =1

3.4. Fusion Decision

D–S evidence theory is a major method of reasoning under uncertainty (RUU). The synthesis
rules of evidences m1 , m2 , . . . , mn are show as Formula (3) for the recognition framework Θ.
Sustainability 2018, 10, 147 6 of 13


 0 A=∅
m( A) = −1 (3)
 (1 − K ) ∑ ∏ m j ( Ai ) A 6= ∅
∩ Ai = A i ≤ j ≤ n

Among them, the basic probability assignment (BPA) satisfies ∑ m j ( Ai ) = 1, and K = ∑ ∑ m j ( Ai )

i ∩ Ai =∅ i
means the degree of conflict between the evidences. Considering the shortcomings of D–S theory mentioned
in Section 2.3, we adopt credibility to modify the evidence body and the local conflict allocation to synthesise
the rules.

1. Evidence body modification based on credibility. The evidence credibility µi (µi ≤ 1) points to
the relative reliability among evidences. The highest reliability of the evidence in the evidence
set was set to 1. µi as a correction factor modifies the original evidence body M0 . The surplus
probability is arranged to the unknown condition m(Θ). The BPA value of the corrected body M
is obtained, as shown in Formula (4):

M = [m( A1 )m( A2 ) . . . m( An )m(Θ)] (4)

n
Among these, m( Am ) = { um0 ( Am )|m = 1, 2, . . . , n}, m(Θ) = 1 − ∑ m( Am ), m0 ( Am ) ∈ M0 .
m =1
2. Synthesis rules based on local conflict allocation. This divides local conflicts between the two
focal points that cause the conflict:

∑

 m jk ( A) =
 m j ( B)mk (C ) + f ( A)
B∩C = A
m2 ( A ) m k ( D ) m2k ( A)m j ( D ) (5)
 f ( A) =
 ∑ ( m (j A)+m + )
j k (D) m j ( D )+mk ( A)
A ∩ D =∅

Among them, B, C, D ⊂ { A1 , A2 , . . . , Θ}, the BPA value satisfies m j (•) ⊂ M j , mk (•) ⊂ Mk ; m jk ( A)

was the BPA value of the new-found evidence body M jk after fusion.
After every two evidences are synthesized, the new evidences carry through normalization.
For example, the element m0jk ( A) is obtained after normalization of Mjk :

m jk ( A)
m0jk ( A) = (6)
∑ m jk (•)

In this paper, location information, the review of the social platform app, product information
and consumer information are set as evidences. We can use statistical analysis or expert experience to
define their credibility.

3.5. Result
The system produces globally consistent user- or resource-characteristic information through
the above information fusion, which will be enriched and changed with time. After a period of
accumulation of the information, user- and resource-feature information is relatively stable. However,
some of the parameters may still be missed due to the instability of multi-source information, so the
system uses a recommendation algorithm based on a power-spectrum estimation. The detailed
algorithms are as follows:
Φ xx (m) is the auto-correlation function of user demand and characteristic information x (n).
Pxx (m) is used to express the power spectral density. The formula for the calculation is as follows:
∞
p xx (ω ) = ∑ Φ xx (m)e−zwm (7)
m=−∞
Sustainability 2018, 10, 147 7 of 13

According to the calculation of power spectral density, Φ xx (m) is denoted as:

N
1
Φ xx (m) = lim ∑
N →∞ 2N + 1 n=− N
x (n) x ∗ (n + m) (8)

After Formula (8) is replaced by Formula (7), there is:

∞
" # " #
N
1
Pxx (ω ) = lim
N →∞ 2N + 1
∑ x (n)e −zwn
• ∑ ∗
x (n + m)e zw(n+m)
(9)
n=− N m=−∞

The average of Formula (9) is:


 2 
1  N 
Pxx (ω ) = lim E  ∑ x (n)ezwn   (10)
 
N →∞ 2N + 1 n=− N 

Because of the multi-sources of user information, there are many elements of user requirements
and feature information, which means the mobile e-commerce recommendation system can only use the
sub-sequence, not the complete information. Therefore, the data-processing module for a recommendation
calculates the power spectral density with a finite number of data sequences. The finite data sequence
of a user’s demand and characteristic information is x(0), x(1), x(2), . . . , x( N ). Calculating the above
formulas, we can obtain:
1 ∞
Φ̂ xx (m) = ∑ x N (n) x N (n + m)
N n=−
(11)
∞

P̂xx (ω ) = ∑ Φ̂xx(m)e−zwm (12)

m

We plug Formula (11) into Formula (12) to generate or modify the recommendation list.

4. Experimental Study
In this paper, the raw data is taken from an online shop selling women’s clothing, Taobao.com.
We selected 100 customers as the research object. The multi-source information in the mobile
e-commerce recommendation system included the following four aspects: location-service information,
commodity-picture information, social-platform information, and the data from the store trading
database. We collected the information. The first step was data reduction, which deleted the data that
did not satisfy the requirements. Consumer age, gender and consumption levels were reserved on
the shop’s transaction database. Then, the data was converted. Due to the limitations of experimental
conditions, collected images were manually converted to keywords. Finally, we conducted unified
coding based on the previous data.
Here is a simple example of the basic process of the recommendation. Suppose we want to collect
relevant information from the database, as shown in Table 1. There is the actual information of consumer A.
The proposed approach collects information from multiple sources, that is, it uses all of the above
information. The traditional method doesn’t collect the information from the location service, pictures
and social platforms. It only needs the data from the store trading database. The weights of different
information sources shown in Table 2 can be determined by utilizing the method proposed in Section 3.3.
Finally, we can use the improved D–S evidence theory to obtain the final recommendation results.
The probability of recommending shoes is as high as 85% according to the proposed method, but the
probability of the traditional method is only 12%. We find that the data sources of recommendation
systems have a great impact on the results. In general, the more the data, the better the recommendation.
For example, we know that consumer A is more active on social networks and assume she has an outgoing
personality. So the types of recommended shoes are young and cute.
Sustainability 2018, 10, 147 8 of 13

Table 1. The information from different platforms about customer A.

Information Location Picture

Shop’s Transaction Database Social Platform
Platform Information Information
Features Age Gender Consumption level Address Picture features Activity Browsing history
A 26 Woman Medium Market Shoes Active Shoes

Table 2. The weights of different information sources between the proposed method and the traditional method.

Weight 26 Woman Medium Market Shoes Active Shoes Other

Proposed method 0.21 0.37 0.36 0 0 0 0 0.06
Traditional method 0.15 0.18 0.16 0.11 0.13 0.04 0.08 0

In the following testing experiments, we test the efficiency of the proposed method in the
recommendation system using the multi-source data by comparing it with traditional recommendation
methods. The comparison items include: recommendation accuracy, simplicity, coverage rate, recall
rate and speed.

4.1. Recommendation Accuracy and Simplicity

Compared with the computer terminal, the mobile terminal visual interface is small, so the
accuracy of the recommendation and the simplicity of the rule become more important for the mobile
e-commerce recommendation system. The simplicity of rules indicate the number of attributes included
in rules. Recommendation accuracy is also very important to the e-tailer. If an e-tailer recommends a
product that does not meet the needs of the consumer, the recommendation system not only cannot
provide personalized decision support and information services for the customers’ shopping, but will
make the consumer feel bored. The comparison of two algorithms is shown in Table 3.

Table 3. Comparison experiment between the traditional algorithm and the proposed algorithm.

Algorithm Accuracy (%) Simplicity

Traditional algorithm 80.01 2.01
Proposed algorithm 91.23 1.47

From Table 3, we can see that the proposed method is more accurate than the traditional algorithm.
This reflects the use of multi-source data of consumers’ behavior to analyze their demands in a more
accurate way. It improves the quality of the recommendation rules.
Due to the diversity of consumer preferences, there are often various possibilities in the mobile
terminal personalized recommendation process. The more products you recommend, the greater
the selection pressure you burden consumers with. So, simplicity is an important indicator of a
recommendation rule’s ability. The mobile terminal recommendation system should provide the most
concise rules that can meet consumers’ different demands. We can see from Table 1 that, in terms of
simple rules, the proposed algorithm is superior to the traditional algorithm, while its recommendation
ability is better than the traditional algorithm. All these prove that the proposed algorithm is significant
in the research of a mobile e-commerce recommendation system. Because of a mobile terminal’s
unlimited accessibility and the diversity of consumer demand, improving the ability to recommend is
very important for ameliorating service quality of the mobile e-commerce recommendation.

4.2. Coverage Rate

The user coverage rate is:

Coverage(t) = ∑ ( RS(t) ∩ RR(t))/∑ RS(t) (13)

SustainabilitySustainability
2018, 10, 147 2018, 10, 147 9 of 13 9 of 13

Coverage(t ) =  ( RS (t ) ∩ RR(t )) /  RS (t ) (13)

where RS(t) is a set of recommended resources and RR(t) is a set of selected resources produced in
where RS (t ) system
the recommendation is a set ofat
recommended
t time. Theresources
higher the RR(t ) is a rate,
and coverage set of the
selected resources
stronger theproduced
coverage on user
in the recommendation
Sustainability 2018, 10, 147 system at t time. The higher the coverage rate,
demands and features in the recommendation system. The user coverage rate of the proposedthe stronger the coverage
9 ofon
13 method
user demands and features in the recommendation system. The user coverage rate of the proposed
is proved in Figure 3.
Coverage(t ) =  ( RS (t ) ∩ RR(t )) /  RS (t )
method is proved in Figure 3.
(13)

where RS (t ) is a set
14 of recommended resources and RR(t ) is a set of selected resources produced
12 system at t time. The higher the coverage rate, the stronger the coverage on
in the recommendation (%)coverage rate (%)
user demands and features in the recommendation system. The user coverage rate of the proposed
10
method is proved in Figure 3.
8

146 traditional algorithm

124 proposed algorithm

User
User coverage rate

102
80
0 5 10 15
6 traditional algorithm
Data time
4 proposed algorithm
2
Figure 3. Comparison experiment of user coverage rate.
Figure
0 3. Comparison experiment of user coverage rate.
Figure 3 shows 0that with 5 continuous 10 enrichment 15 of the user information in the two
recommender systems, the user discovery Data timecoverage rates increase continuously until they reach their
Figurerespective
3 showspeaks thatorwith continuous enrichment of the user information in the two recommender
close to them. Furthermore, the proposed algorithm of the new system is more
systems, the user discovery
optimized. So, from thecoverage
perspectiverates
of the increase
user coveragecontinuously until they
speed, the advantages reach
of the new their
systemrespective
Figure 3. Comparison experiment of user coverage rate.
peaks or close to them.
are obvious, Furthermore,
particularly the proposed
when it comes to the rapidalgorithm
change of in ofthethe newofsystem
number consumersis more
becauseoptimized.
So, from theof corporate
perspective advertising,
of the promotions and other activities. Therefore, judging thefrom thesystem
rate of the
Figure 3 shows thatuser
withcoverage
continuousspeed, the advantages
enrichment of the user of new
information in the two are obvious,
user discovery coverage, the advantages of the new system are obvious, especially when there is a
particularlyrecommender
when it systems,tothe
comes theuser discovery
rapid changecoverage
of rates
in the increase
number continuously
of until they
consumers reach their
because of corporate
sharp increase of the number of consumers in large-scale promotional activities.
respective peaks or close to them. Furthermore, the proposed algorithm of the new system is more
advertising, promotions and other activities. Therefore, judging from the rate of the user discovery
optimized. So, from the perspective of the user coverage speed, the advantages of the new system
4.3.
coverage, are Recall
theobvious, Rate
advantages of the newit system
particularly when comes to are obvious,
the rapid changeespecially whenofthere
of in the number is a sharp
consumers becauseincrease of
the number of of consumers
The recall
corporate rate in large-scale
refers
advertising, promotions promotional
to that number
andof selected
other activities.
resources
activities. from recommendation
Therefore, judging from thesystems
rate of that
the
accounts
user for thecoverage,
discovery total number of selectedofresources
the advantages the new selected
system areby obvious,
the user. especially
CA: 20 customers.
when thereCB:is40
a
customers.
4.3. Recall sharp
Rate increase CC:of60the
customers.
number ofCD: 80 customers.
consumers TF: 100promotional
in large-scale customers. The results of the comparison
activities.
test are shown in Figure 4.
The recall rate
4.3. Recall refers to that number of selected resources from recommendation systems that
Rate
accounts for the
The total
recall number
rate refers8of selected
to that number resources
of selectedselected
resources by
fromthe user. CA: 20systems
recommendation customers.
that CB: 40
The recommendation

customers. CC: 60 customers. CD:

accounts for the total number
6 80 customers. TF: 100 customers. The results of the comparison test
of selected resources selected by the user. CA: 20 customers. CB: 40
recall rate(%) recall rate(%)

are showncustomers.
in FigureCC:4. 60 customers.
4
CD: 80 customers. TF: 100 customers. The results of the comparison
test are shown in Figure 4.
2
80
The recommendation

6 CA CB CC CD CE
User set
4

2 traditional algorithm proposed algorithm

0
CA
Figure 4. Comparison CB of the recommendation
test CC CD recall rate.
CE
User set

traditional algorithm proposed algorithm

Figure 4. Comparison test of the recommendation recall rate.

Figure 4. Comparison test of the recommendation recall rate.

From Figure 4, the recall rates of the two systems gradually show a steady state with the increasing
abundance of users and resource information. Also, because of the steadiness of user demands this starts
to drop. However, the multi-source information-fusion method in the new system makes the tendency
more obvious in the recall rate, and the system maintains a high recall rate over a period of time.
 From Figure 4, the recall rates of the two systems gradually show a steady state with the
increasing abundance of users and resource information. Also, because of the steadiness of user
demands this starts to drop. However, the multi-source information-fusion method in the new
system makes
Sustainability 2018, the tendency more obvious in the recall rate, and the system maintains a high10recall
10, 147 of 13
rate over a period of time.

4.4.
4.4. Recommendation
Recommendation Speed
Speed
The
The recommendation
recommendation speed speed is is the
the time
time from
from the
the input
input ofof the
the data
data to
to the
the output
output of of the
the
recommended product. In the process of multi-source information collection
recommended product. In the process of multi-source information collection and conversion, we and conversion,
we deducted
deducted the the time
time of ofmanual
manualconversion.
conversion. A Acomparison
comparisonexperiment
experiment ofof aa multi-source
multi-source
recommendation
recommendationsystem systemand a traditional
and recommendation
a traditional system system
recommendation followed. The traditional
followed. algorithm
The traditional
normally only scans a shop’s transaction database, but the article proposes a
algorithm normally only scans a shop’s transaction database, but the article proposes a methodmethod integrating
multi-source information information
integrating multi-source to recommended commodities.commodities.
to recommended The experiment Thedemonstrated that the time
experiment demonstrated
taken
that the time taken for the proposed method in each group of transactions is longer thantraditional
for the proposed method in each group of transactions is longer than under the under the
algorithm.
traditional As volumes As
algorithm. built up, the built
volumes difference became
up, the greater.
difference The times
became taken
greater. fortimes
The the two algorithms
taken for the
are
twopresented
algorithms inare
Table 4 and Figure
presented 5. 4 and Figure 5.
in Table

Table 4.
Table Comparison experiment
4. Comparison experiment of
of operating
operating efficiency,
efficiency, according
according to
tothe
thenumber
numberof
ofusers.
users.

Users UsersTraditional Algorithm

Traditional (s) (s)
Algorithm ProposedAlgorithm
Proposed Algorithm(s) (s)
CA CA 0.33 0.33 0.41
0.41
CB CB 0.65 0.65 0.83
0.83
CC CC 0.96 0.96 1.32
1.32
CD CD 1.28 1.28 1.58
1.58
CE 1.41 1.99
CE 1.41 1.99

2.5
2
Time (s)

1.5
1
0.5
0
CA CB CC CD CE
Number of users
Traditional algorithm (s) Proposed algorithm (s)

Figure 5. A comparison of the two algorithms, according to the number of users.

Figure 5. A comparison of the two algorithms, according to the number of users.

We use the above experiments to compare the time difference registered between the proposed
We use
algorithm thethe
and above experiments
traditional to compare
algorithm the time
for different difference
amounts registered
of data. betweenshow
Experiments the proposed
that the
algorithm
proposed algorithm is a time-consuming operation. In other words, the improved algorithmthe
and the traditional algorithm for different amounts of data. Experiments show that is
proposed
unable to algorithm
enhance theis aspeed
time-consuming operation.system;
of a recommendation In otherbecause
words, the improved
of the algorithm
information is unable
diversification
to enhance
of the the
mobile speed of the
consumer, a recommendation
improved algorithm system;
takesbecause
longer. of the provide
It can information diversification
useful information forof
the mobile consumer,
consumers, the improved
but also provide rules foralgorithm takes longer.
the decision-making It can
ability provide
of the useful information for
enterprise.
consumers, but also provide rules for the decision-making ability of the enterprise.
5. Conclusions and Discussion
5. Conclusions and Discussion
This paper shows that a mobile e-commerce recommendation system based on the proposed
This paper shows that a mobile e-commerce recommendation system based on the proposed
algorithm can provide consumers with the most necessary information at the most appropriate time,
algorithm can provide consumers with the most necessary information at the most appropriate time,
and provide a more comfortable shopping experience for them. Current literature regards
and provide a more comfortable shopping experience for them. Current literature regards
recommendation systems that only collect information from the shopping platform. According to
recommendation systems that only collect information from the shopping platform. According to
the characteristics of the mobile e-commerce recommendation system, the proposed system
the characteristics of the mobile e-commerce recommendation system, the proposed system collected
consumers’ requirements from several approaches. It can provide consumers with more accurate
recommendations. Accurate recommendations can not only enhance customer satisfaction, but also
affect the conversion rate of goods. Thus, e-tailers are advised to recommend accurately the most
suitable products in order to boost sales. E-trailers should take advantage of a user’s multi-source
Sustainability 2018, 10, 147 11 of 13

information so as to guide and cultivate consumers’ decision-making by combining the user’s

mobile-phone shopping habits. They can implement precise recommendations through research
on user demand, consumer behavior and consumption preference. So, this research enriches existing
theory and helps improve managerial practice. At the same time, this paper is of great significance for
mobile e-commerce platform providers in the design of a product recommendation system. Mobile
e-commerce platform operators could re-design the algorithms of their recommendation systems based
on the above in order to choose more suitable recommended products. This addresses the specialities
of a mobile platform compared to a personal computer (PC) and develops a useful tool for targeting
consumers with personalized messages.
Although the article has put forward multi-source information fusion technology to improve
the prediction accuracy of personalized recommendations in the mobile e-commerce terminal, some
restrictions also need to be pointed out. On the one hand, the experimental results in Section 4.4
show that the computation time of the proposed algorithm is longer than the traditional algorithm.
In terms of speed, the proposed algorithm has no advantages. So we will consider how to ameliorate
the proposed method in terms of its speed in future research. On the other hand, in the process of
designing a recommendation system, we only focus on a user’s convenience regarding how to find the
products of interest to them precisely. The recommendation system could also affect consumer privacy.
Customer information security issues have become the main bottleneck that restricts the growth of
mobile commerce.

Acknowledgments: The work was supported by the soft science research project of Chengdu city (Project
No. 2015-RK00-00296-ZF).
Author Contributions: Mingfu Li built the framework of the whole paper. Chengxin Yin carried out the collection
and sorting of the data. Xiaoting Ren designed the whole experiment and method. Ping Liu provided analytical
and experimental tools. Yan Guo finished all the experiments and the whole article.
Conflicts of Interest: The authors claim that there is no conflict of interest.

References
1. Martín, S.S.; López-Catalán, B.; Ramón-Jerónimo, M.A. Factors determining firms’ perceived performance of
mobile commerce. Ind. Manag. Data Syst. 2012, 112, 946–963. [CrossRef]
2. Guo, Y.; Wang, M.X.; Li, X. Application of an improved Apriori algorithm in a mobile e-commerce
recommendation system. Ind. Manag. Data Syst. 2017, 117, 287–303. [CrossRef]
3. Ansari, A.; Essegaier, S.; Kohli, R. Internet recommendation systems. J. Market. Res. 2000, 37, 363–375.
[CrossRef]
4. Bodapati, A.V. Recommendation systems with purchase data. J. Market. Res. 2008, 45, 77–93. [CrossRef]
5. Jiang, Y.C.; Shang, J.; Liu, Y.Z. Maximizing customer satisfaction through an online recommendation system:
A novel associative classification model. Decis. Support Syst. 2010, 48, 470–479. [CrossRef]
6. Zhao, Y.S.; Liu, Y.P.; Zeng, Q.A. A weight-based item recommendation approach for electronic commerce
systems. Electron. Commer. Res. 2015, 19, 1–22. [CrossRef]
7. Zenebe, A.; Norcio, A.F. Representation, similarity measures and aggregation methods using fuzzy sets for
content-based recommender systems. Fuzzy Sets Syst. 2009, 160, 76–94. [CrossRef]
8. Lin, Z.J. An empirical investigation of user and system recommendations in e-commerce. Decis. Support Syst.
2014, 68, 111–124. [CrossRef]
9. Herlocker, J.L.; Konstan, J.A.; Loren, J.; Terveen, G.; Riedl, T. Collaborative filtering recommender systems.
ACM Trans. Inf. Syst. 2004, 22, 5–53. [CrossRef]
10. Liu, Y.Z.; Jiang, Y.C.; Liu, X.; Yang, S.L. CSMC: A combination strategy for multi-class classification based on
multiple association rules. Knowl.-Based Syst. 2008, 21, 786–793. [CrossRef]
11. Cho, Y.H.; Kim, J.K.; Kim, S.H. A personalized recommender system based on web usage mining and
decision tree induction. Expert Syst. Appl. 2002, 23, 329–342. [CrossRef]
12. Rak, R.; Kurgan, L.; Reformat, M. A tree-projection-based algorithm for multi-label recurrent-item
associative-classification rule generation. Data Knowl. Eng. 2008, 64, 171–197. [CrossRef]
Sustainability 2018, 10, 147 12 of 13

13. Wang, F.H.; Shao, H.M. Effective personalized recommendation based on time-framed navigation clustering
and association mining. Expert Syst. Appl. 2004, 27, 365–377. [CrossRef]
14. Liang, T.P.; Yang, Y.F.; Chen, D.N.; Ku, Y.C. A semantic-expansion approach to personalized knowledge
recommendation. Decis. Support Syst. 2008, 45, 401–412. [CrossRef]
15. Choi, K.; Yoo, D.; Kim, G.; Suh, Y. A hybrid online-product recommendation system: Combining implicit
rating-based collaborative filtering and sequential pattern analysis. Electron. Commer. Res. Appl. 2012, 11,
309–317. [CrossRef]
16. Mikalef, P.; Giannakos, M.; Pateli, A. Shopping and word-of-mouth intentions on social media. J. Theor. Appl.
Electron. Commer. Res. 2013, 8, 17–34. [CrossRef]
17. Hsiao, K.L.; Lin, J.C.; Wang, X.Y.; Lu, H.P.; Yu, H. Antecedents and consequences of trust in online product
recommendations: An empirical study in social shopping. Online Inf. Rev. 2010, 34, 935–953. [CrossRef]
18. Guo, Y.; Wang, M.X.; Li, X. An interactive personalized recommendation system using the hybrid algorithm
model. Symmetry 2017, 9, 216. [CrossRef]
19. Yuto, I.; Takahiro, U.; Ichi, T. Design and evaluation of a movie recommendation system showing a review
for evoking interested. Int. J. Web Inf. Syst. 2017, 13, 72–84.
20. Shani, G.; Gunawardana, A. Evaluating recommendation systems. In Recommender Systems Handbook;
Springer: Berlin, Germany, 2011; pp. 257–297.
21. Benbasat, I.; Wang, W. Trust in and adoption of online recommendation agents. J. Assoc. Inf. Syst. 2005, 6, 4.
22. Goldberg, D.; Nichols, D.; Oki, B.M.; Terry, D. Using collaborative filtering to weave an information tapestry.
Commun. ACM 1992, 35, 61–70. [CrossRef]
23. Masoud, M.; Mehdi, S. Improving recommender systems’ performance on cold-start users and controversial
items by a new similarity model. Int. J. Web Inf. Syst. 2016, 12, 126–149.
24. Pazzani, M.J.; Billsus, D. Content-based recommendation systems. In The Adaptive Web; Springer:
Berlin/Heidelberg, Germany, 2007; pp. 325–341.
25. Cheung, K.W.; Kwok, J.T.; Law, M.H.; Tsui, K.C. Mining customer product ratings for personalized marketing.
Decis. Support Syst. 2003, 35, 231–243. [CrossRef]
26. Cheng, L.C.; Jhang, M.J. A novel approach to exploring maximum consensus graphs from users’ preference
data in a new age environment. Electron. Commer. Res. 2015, 15, 543–569. [CrossRef]
27. Scholz, M.; Dorner, V.; Franz, M.; Hinz, O. Measuring consumers’ willingness to pay with utility-based
recommendation systems. Decis. Support Syst. 2015, 72, 60–71. [CrossRef]
28. Xiao, B.; Benbasat, I. E-Commerce product recommendation agents: Use, characteristics, and impact. MIS Q.
2007, 31, 137–209. [CrossRef]
29. McDonald, D.W.; Ackerman, M.S. Expertise recommender: A flexible recommendation system and
architecture. In Proceedings of the 2000 ACM Conference on Computer Supported Cooperative Work,
Philadelphia, PA, USA, 2–6 December 2000; pp. 231–240.
30. Yager, R.R. Some considerations in multi-source data fusion. Stud. Comput. Intell. 2005, 5, 1–22.
31. Pavlin, G.; de Oude, P.; Maris, M.; Nunnink, J.; Hood, T. A multi-agent Systems Approach to Distributed
Bayesian Information Fusion. Inf. Fusion 2010, 11, 267–282. [CrossRef]
32. Villegas, D.; Bobroff, N.; Rodero, I.; Delgado, J.; Liu, Y.; Devarakonda, A.; Fong, L.; Sadjadi, S.M.; Parashar, M.
Cloud federation in a layered service mode. J. Comput. Syst. Sci. 2012, 78, 1330–1344. [CrossRef]
33. Li, X.Y.; Zhou, F.; Yang, X.D. Developing Dynamic P2P Trust Model Using Theory of Entropy-based
Multi-source Information Fusion. Int. J. Innov. Comput. Inf. Control 2011, 7, 7–790.
34. Su, H.S. Multi-source Fuzzy Information Fusion Method Based on Bayesian Optimal Classifier. Acta Autom. Sin.
2008, 34, 282–287. [CrossRef]
35. Yun, Z.; Quan, Z.; Caixin, S.; Shaolan, L.; Yuming, L.; Yang, S. RBF neural network and ANFIS-based
short-term load forecasting approach in real-time price environment. IEEE Trans. Power Syst. 2008, 23,
853–858.
36. Shafer, G. A Mathematical Theory of Evidence; Princeton University Press: Princeton, NJ, USA, 1976.
37. Bi, T.S.; Ni, Y.X.; Wu, F.L.; Yang, Q.X. Hybrid fault section estimation system with radial basis function neural
network and fuzzy system. Proc. CSEE 2005, 25, 12–18.
38. Yingwei, L.; Sundararajan, N.; Saratchandran, P. Performance evaluation of a sequential minimal radial basis
function (RBF) neural network learning algorithm. IEEE Trans. Neural Netw. 1998, 9, 308–318. [CrossRef]
[PubMed]
Sustainability 2018, 10, 147 13 of 13

39. Huang, G.B.; Saratchandran, P.; Sundararajan, N. A generalized growing and pruning RBF (GGAP-RBF)
neural network for function approximation. IEEE Trans. Neural Netw. 2005, 16, 57–67. [CrossRef] [PubMed]
40. Anand, S.S.; Bell, D.A.; Hughes, J.G. EDM: A general framework for Data Mining based on Evidence Theory.
Data Knowl. Eng. 1996, 18, 189–223. [CrossRef]
41. Skowron, A.; Grzymala-Busse, J. From Rough Set Theory to Evidence Theory. In Advances in the Dempster-Shafer
Theory of Evidence; Yaeger, R., Fedrizzi, M., Kacprzyk, J., Eds.; John Wiley & Sons, Inc.: New York, NY, USA,
1994; pp. 193–236.
42. Yang, S.L.; Liu, Y.Z.; Li, Y.F. Evidence acquisition and synthesis method based on theory. J. Manag. Sci. 2005,
8, 69–75.
43. Deng, Y.; Shi, W.; Zhu, Z.; Liu, Q. Combining belief functions based on distance of evidence. Decis. Support Syst.
2004, 38, 489–493.
44. Lang, F.H.; Gu, L.Z.; Yang, Y.X.; Niu, X.X. An improved combination rule of conflict evidence. J. Beijing Univ.
Posts Telecommun. 2008, 31, 84–88.
45. Murphy, C.K. Combining belief functions when evidence conflicts. Decis. Support Syst. 2000, 29, 1–9. [CrossRef]
46. Guo, H.W.; Shi, W.K.; Liu, Q.K.; Deng, Y. A new combination rule of evidence. J. Shanghai Jiaotong Univ. 2006,
40, 1895–1900.
47. Li, P.; Yan, Y.; Zhang, W.J.; Jia, J.F.; Bai, B. Short-term load forecasting based on quantum neural network by
evidential theory. Power Syst. Prot. Control 2010, 38, 49–53.
48. Zhang, Z.G.; Yang, Y.H.; Xia, L.X.; Jia, Y.H. The application of RPROP algorithm to igneous rock classification.
Prog. Geophys. 2008, 23, 898–902.
49. Jiang, H.L.; Guan, Y.; Liu, F.; Xu, J. Transmission line fault diagnosis based on RBF network with HCM
clustering algorithm. High Volt. Eng. 2005, 31, 81–84.

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).