Conference PaperPDF Available

Discovering Spatio-Temporal Relationships among IoT Services

Authors:

Abstract and Figures

We propose a framework to discover proximate IoT service relationships based on spatio-temporal features. We introduce a spatio-temporal proximity model in terms of spatial-proximity and temporal-proximity to discard insignificant IoT service relationships. The proximity model focuses on quantifying the correlation strength among IoT services from time and location aspects. A new algorithm is proposed to discover proximate spatio-temporal IoT service relationships. We also present preliminary experimental results.
Content may be subject to copyright.
Discovering Spatio-temporal Relationships among IoT Services
Bing Huang, Athman Bouguettaya, Azadeh Ghari Neiat
School of Information Technology,
University of Sydney, Australia
{bing.huang, athman.bouguettaya, azadeh.gharineiat }@sydney.edu.au
Abstract—We propose a framework to discover proximate
IoT service relationships based on spatio-temporal features.
We introduce a spatio-temporal proximity model in terms of
spatial-proximity and temporal-proximity to discard insignificant
IoT service relationships. The proximity model focuses on
quantifying the correlation strength among IoT services from
time and location aspects. A new algorithm is proposed to
discover proximate spatio-temporal IoT service relationships.
Keywords-spatio-temporal relationships; Internet of Things;
proximity; spatio-temporal data;
I. INT ROD UC TI ON
The Internet of Things (IoT) paradigm enables phys-
ical devices to connect and exchange information. IoT
devices allow objects to be sensed or controlled remotely
through Internet [1]. IoT devices are highly heterogeneous in
terms of supporting infrastructure ranging from networking
to programming abstraction[?]. Service-oriented Computing
(SOC) is a promising solution for abstracting things on
the Internet as services by hiding the complex and diverse
supporting infrastructure [3]. This abstraction can shift the
focus from dealing with technical details to how services
are to be used. We refer services for Internet of Things as
IoT services. For example, a light which is connected to the
Internet can be represented as a light service.
Spatio-temporal features are important for IoT services
because they characterize the time and the location of
IoT services’ usage. In this paper, we focus on discover-
ing spatio-temporal relationships among IoT services (i.e.,
we refer such spatio-temporally correlated IoT services as
composite1IoT services). We also use the terms composite
IoT services and spatio-temporally correlated IoT services
interchangeably. It is important to discover spatio-temporal
IoT service relationships. For example, the light, the TV,
and the air-conditioning in the bedroom may be turned on
in the evening as shown in Figure 1. The spatio-temporal
relationship among these IoT services may provide valuable
information to help residents better understand their IoT
service usage patterns. These patterns can be used to design
intelligent control of IoT services in smart homes to reduce
residents’ interactions with IoT services.
1We employ composite” and “composed interchangeably.
Discovering useful IoT service relationships is challeng-
ing because there are many possibilities of establishing
spatio-temporal relationships among IoT services. Too many
IoT service relationships may hinder residents from under-
standing their actual IoT service usage behaviors. To address
this challenge, there is a need to design a technique to filter
out insignificant spatio-temporal IoT service relationships.
In this paper, we use the proximity among component IoT
services to indicate the significance of the discovered spatio-
temporal relationships. The proximity of the composite IoT
services characterizes the correlation strength of spatio-
temporal relationships among component IoT services.
The usage history of IoT services is recorded in IoT
service event sequences. A snapshot of IoT service event
sequences and their pictorial representation are shown in
Figure 3 and Figure 2, respectively. Proximate IoT service
relationships are implicit and hidden in these sequences. In
this paper, we focus on the problem of discovering proxi-
mate spatio-temporal IoT service relationships from service
event sequences. The key contributions of this paper are as
follows: (1) A new IoT service model is proposed based
on spatio-temporal features. We also introduce the notion
of composite IoT services to capture the spatio-temporal
relationships among IoT services. (2) A spatio-temporal
proximity model is introduced to capture correlation strength
of composite IoT services in terms of spatial-proximity
and temporal-proximity. We develop an efficient algorithm
called P M iner (Proximate Composite IoT service Miner)
to discover proximate composite IoT services.
s1
s3
Relative Time
AC
Light
TV
0 20 40 60 80
s2
Composite IoT
Service
Light
Shower
Toilet Sink
Fridge
Microwave
Toaster
Light
TV
Resident 1
Resident 2
Resident Part of discovered composite IoT sevices
Figure 1. An example of composite IoT service
0 20 40 60 80 100
Bedroom
Living room
Bedroom
Bathroom
A
D
B
C
Day 1
F
E
Day 2
Day 3
M
Relative Time
Clock
LED
Heater
Ventilator
Bathroom Music player
Bathroom Shower
Bedroom
Living room
Bedroom
A
B
C
Clock
LED
Heater
Gym
Gym Airconditioner
KTreadmill
Bathroom EMusic player
Bathroom FShower
Bedroom
Living room
Bedroom
A
B
C
Clock
LED
Heater
G
Living room TV
Bathroom FShower
Location ...
Figure 2. Pictoral representation of service event sequences
Day 1
A+
A-
B+
C+
D+
C-
B-
D-
E+
F+
F-
E-
0
1
5
6
10
22
23
30
60
61
73
75
Day 2
A+
A-
B+
M+
C+
C-
B-
K+
M-
K-
E+
F+
F-
E-
0
1
4
5
7
23
24
30
40
60
70
72
82
84
Day 3
A+
A-
B+
C+
C-
B-
G+
G-
F+
F-
0
1
6
10
25
26
40
60
64
77
Figure 3. An example of service event sequences
II. MODELING
We first introduce the notion of IoT services based on
spatio-temporal features [2]. Then we propose the notion
of composite IoT services to capture spatio-temporal rela-
tionships among component IoT services. Lastly, a spatio-
temporal proximity model is proposed to quantify the cor-
relation strength of the composite IoT services.
A. Spatio-temporal model for IoT service
Definition 1: IoT Service. An IoT service Siis a tuple
Si=< si, Fi, IS, F S >, where:
siis a unique service identifier.
Fiis a set of functionalities that are offered by Si.
IS (Initial State) is a tuple < s+
i, ls, sti>, where
s+
iis a symbol of IS.
ls=< xs, ys, zs>is a start location of Si, where
< xs, ys, zs>is a GPS point.
stiis a start-time of Si.
F S (Final State) is a tuple < s
i, le, eti>, where
s
iis a symbol of F S
le=< xe, ye, ze>is an end location of Si, where
< xe, ye, ze>is a GPS point.
etiis an end-time of Si.
We assume the start location and the end location
are the same during IoT service functionality deliver-
ing and can be detected in smart homes. Since we
focus on the spatio-temporal features in the remainder
of this paper, the representation of the IoT service Si
can also be simplified as < si,(s+
i, sti),(s
i, eti), li>.
For example, a light service can be represented as <
light, (light+,7pm),(light,9pm),(1,2,0) >.
B. Spatio-temporal model for composite IoT service
Multiple IoT services may be used collectively based on
time and location correlations. Spatio-temporal correlated
IoT services are referred as a composite IoT service.
Definition 2: Composite IoT Service. A composite IoT
service is denoted by a tuple CS =< S, sup > where
S={< s1,(s+
1, st1),(s
1, et1), l1>, ..., <
sn,(s+
n, stn),(s
n, etn), ln>}represents ncomponent
IoT services, where stisti+1 and stieti.
An example of composite IoT service is shown in
Figure 1. For brevity, we omit the device ID si
and sort all elements s
i(* can be + or -) in S
in a non-decreasing order based on its associated
time information sti(or eti). Scan be represented
as S=< Seq , T, L >=
α1... αi... α2n
t1... ti... t2n
l1... li... l2n
,
where Seq ={α1...αi...α2n}is the state symbol
sequence and αi=s
j(* can be + or -),
T={t1...ti...t2n}is the time and titi+1,
and L={l1...li...l2n}is the location.
For example, the composite IoT service in Figure 1 can
be represented as
s+
2s+
1s+
3s
1s
3s
2
48 50 58 65 70 75
l2l1l3l1l3l2
(i.e., l1=<1,2,3>,l2=<2,4,3>,l3=<3,5,1>)
sup is the support for S. The support sup is a measure
reflecting the usefulness of the composite IoT service
[4]. Let S=< Seq, T , L > and S0=< Seq0, T 0, L0>
be two composite IoT services. Sis referred to as a
sub-composite IoT service of S0, denoted as SvS0,
if Seq ={α1...αi...α2n}is a subsequence of S eq0=
{α0
1...α0
i...α0
2m}, denoted as Seq vS eq0.Seq vSeq0
is satisfied if there exists integers 1k1k2...kn
kmsuch that α1α0
k1,α2α0
k2,..., αnα0
kn.
Given a spatio-temporal database DB which records IoT
service event sequences, the tuple (sid, S0)(i.e., sid is a
sequence ID and S0is the composite IoT service) is said
to contain a sub-composite IoT service Sif SvS0. The
support sup of Sin DB, denoted as sup(S)is the number of
tuples containing S.sup(S)can be formalized as follows.
sup(S) = |{(sid, S)DB|SvS0}| (1)
C. Spatio-temporal proximity model for composite IoT
services
A proximity model is proposed in terms of
spatio proximity and temporal proximity to quantify
the spatio-temporal correlation strength among component
IoT services. The proximity model is based on the approach
of measuring spatio-temporal interval data distance used in
clustering spatio-temporal interval data [6].
Definition 3: Proximity. Given a composite IoT service
CS =< S, sup > where Sis in the following representation
form S=< Seq, T, L >=
α1... αi... α2n
t1... ti... t2n
l1... li... l2n
,its
proximity function is defined in Equation (2).
U=w1·spatial proximity +w2·temporal proximity
(2)
where wi[0,1] (i=1, 2) is a weight such that w1+w2= 1.
spatial proximity: The spatial proximity is based on
the spatial aspect. It measures the average location
proximity among all component IoT services. We first
formalize the spatial proximity for one compos-
ite IoT service in Equation (3). Then the average
spatial proximity for the composite IoT service is
formalized in Equation (4).
Spa =
n
X
i=1
1
|xixi+1|+|yiyi+1 |+|zizi+1|
(3)
spatial proximity =Psup
j=1 Spaj
sup (4)
where nis the total number of component IoT services,
li=< xi, yi, zi>and li+1 =< xi+1, yi+1 , zi+1 >
are two locations for two consecutive component IoT
services and sup is the support. For example, for the
composite IoT service in Figure 1, its spatio proximity
score is Spa =1
|l2l1|+1
|l1l3|=1
|21|+|42|+|33|+
1
|13|+|25|+|31|= 0.473.
temporal proximity: The temporal proximity is based
on the temporal aspect. It measures the average tempo-
ral proximity among component IoT services. The idea
of temporal proximity borrows from the technique of
clustering time-interval based data in [6]. For each com-
ponent IoT service Si={si,(s+
i, sti),(s
i, eti), li},
we utilize a function fiwith respect to tto map the
temporal aspect of Si.fiis formalized in Equation (5).
fi(t) = (1, t [sti, eti]
0, otherwise (5)
Then we have a set of functions {f1, f2, ...fn}cor-
responding to the composite IoT service CS . The
temporal proximity for the composite IoT service can
be calculated by Equation (6). The average tempo-
ral proximity is calculated by Equation (7).
T emp =Rt2n
t1Pn
i=1 fi(t)dt
(t2nt1)·n(6)
temporal proximity =Psup
j=1 T empj
sup (7)
where t1and t2nare the first and the last time informa-
tion of CS, respectively, and nis the number of compo-
nent IoT services. For example, the temporal proximity
score of the composite IoT service {<stove, [18:00,
19:00]>,<washing machine, [18:40, 19:20 ] >}can be
calculated as (18:4018:00)+(19:0018:40)·2+(19:2019:00)
(19:2018:00)·2
= 0.625. This composite IoT service can be interpreted
as when a resident is cooking, he is also doing laundry.
III. DISCOVERING PROXI MATE COMPOSITE IOT
SERVICES APPROACH
We develop a new algorithm P Miner to efficiently
discover proximate composite IoT services from IoT service
event sequences. P Miner consists of two phases. First,
P M iner searches all frequent composite IoT service can-
didates from service event sequences. Second, P Miner
calculates proximity for each candidate and discards those
having low proximity. For the sake of consistency with the
terms from data analysis techniques, we use event patterns
to refer to composite IoT services occurring in service event
sequences. We use the running example in Figure 3 to
illustrate the process of P M iner shown in Figure 4.
Phase I: Searching event patterns.PM iner employs an
Apriori-based approach and find-then-index strategy, where
n-length patterns are used to find (n+1)-length patterns
through exploring the index sets associated with the n-length
patterns rather than searching the whole database [4][5].
P M iner searches event patterns in three sub-phases.
1. Find the set of 1-length event patterns.P Miner first
scans the DB to count the number of each event pairs. If
the user given minsup is 2, all discovered 1-length event
patterns whose support is more than 2 constitute the 1-length
event set L1as shown in Figure 4.
2. Construct an index set for each 1-length event pattern. Let
L1={p1, p2, ...pn}be the set of 1-length event patterns.
For each pi(1in), a corresponding index set piidx
is created. piidx is a collection of elements indicating the
supporting sequences for pi. Each element of piidx contains
three fields, namely, sid,time, and location.sid indicates
the sequence containing pi.time is the time information of
pi.location is the location information of pi. For the sake
of brevity, the location information is omitted in the running
example because it remains the same.
Figure 4. The Process of P M iner
3. Leverage Lk1to find Lkthrough searching the index sets
corresponding to Lk1(k2). Let Lk1={ρ1, ρ2, ...ρm}
be the set of (k-1)-length event patterns. For each member
ρiin Lk1(1im), its index set is ρiidx, resulting in
the set Ik1={ρ1idx, ρ2idx, ..., ρnidx}corresponding to
Lk1. To find Lk, a three-step process including join step,
prune step and construct index set step is performed.
The join step: Let the sequences ρi=
{a1, a2, ..., a2k2}and ρj={b1, b2, ..., b2k2}
be two members in Lk1, where aiand biare event
types. The corresponding index sets for ρiand ρjare
ρiidx and ρjidx, respectively. ρiand ρjare potentially
joinable if ρiidx and ρjidx share at least minsup
numbers of sid, that is, W=ρiidx.sid ρjidx.sid =
{sidw1, sidw2, ...sidwm}such that |W| minsup.
This condition can efficiently reduce the generation
of candidates. The join, ρiρj, is performed as
follows. For each shared sidwi W, we can generate
the following sequences, denoted by γwi, by merging
ρi={a1, a2, ..., a2k2}and ρj={b1, b2, ..., b2k2}
and sorting these merged elements based on their
time information. Thus the outcomes of joining is
γ=ρiρj={γw1, γw2, ..., γwm}where γw i is the
new ordered sequence for the shared sequence sidwi.
By performing join action for members in Lk1, the
outcomes constitute the candidate set Ck. Continue
the running example, L1is used to find C2. Every two
members in L1are joined and the outcomes of the
join step constitute C2in Figure 4.
The prune step: Let γwi and γwj be two members in γ.
γwi and γwj are the same if their elements are the same,
that is, (γwi[1] = γw j [1]) (γwi[2] = γw j [2]) ...
(γwi[h] = γw j [h]). For example, < F +FG+G>
and < F +G+FG>are two different members.
The prune step focuses on counting the numbers of the
same members in γand filter out those members whose
numbers are less than minsup. In the running example,
no members in C2are filtered out.
Construct index set: Given γ=ρiρj=
{γw1, γw2, ..., γwm}after the prune step, collect the
entry < sid, time, location > from ρiidx and ρjidx
for each γwi. All entries constitute the index set for γ.
Phase II: Calculate proximity for event patterns. Given
the event patterns generated by the phase I and minimum
proximity threshold minpro, calculate average proximity
for each event pattern by Definition 3 and filter out those
patterns whose proximity are less than minpro. For the
running example, if the weight for spatial proximity and
temporal proximity is set to be 0 and 1, respectively, the
proximity members in C2is P(A+AB+B)=0.418. If
the minpro is set to be 0.3, members in C2whose proximity
is less than 0.3 are filtered out, resulting in the set L2.
By recursively performing the join step, prune step, and
construct index set step, we can find L3. The ultimate out-
comes of the running example are < E+F+FE>and
< A+AB+C+CB>and their respective proximity
scores are 0.879 and 0.494.
IV. CONCLUSION AND FUTURE WORK
This paper propose a novel framework for discovering
proximate spatio-temporal IoT service relationships. We
design an efficient algorithm P M iner to search significant
composite IoT services. Future work focuses on improving
the performance of P M iner and explore semantic anno-
tation techniques to enhance the readability of discovered
composite IoT services.
REFERENCES
[1] Atzori. L, Iera. A, “The internet of things: A survey,” Journal
of Computer networks 54(15), 2787-2805 (2010).
[2] Neiat. A. G, Bouguettaya. A, “Crowdsourced Coverage as a
Service: Two-Level Composition of Sensor Cloud Services,
IEEE Transactions on Knowledge and Data Engineering,
29(7), 1384-1397(2017).
[3] Guinard. D, Trifa. V, “Interacting with the soa-based internet of
things: Discovery, query, selection, and on-demand provision-
ing of web services,” IEEE transactions on Services Computing
3(3), 223-235 (2010).
[4] Han. Jiawei, Jian Pei and Micheline Kamber, “Data mining:
concepts and techniques,” Elsevier, 2011.
[5] Lin. M. Y and Lee. S. Y, “Fast discovery of sequential patterns
by memory indexing, In International Conference on Data
Warehousing and Knowledge Discovery. pp. 150-160(2002).
[6] Shao and Bouguettaya , “Clustering big spatiotemporal-interval
data,” IEEE Transactions on Big Data. pp.190-203 (2016).
... (3) The spatio-temporal constraints of the functionalities provided by IoT services put forward additional considerations for efficient service collaboration [6]. The IoT system is a system with strong spatio-temporal constraints, that is, the perception and control provided by IoT devices need to be done at a specific time and space. ...
... But it does not get more in-depth with service modeling and service invocation. Huang et al. [6] proposes a framework to discover the spatio-temporal relationships between IoT services, which can help with accurate collaboration between IoT services. These works provide important ideas for efficient spatio-temporal constraints in dynamic IoT systems. ...
Article
Realizing the dynamic collaboration between Internet of Things (IoT) services is the key to developing intelligent IoT applications. However, heterogeneity, changeable invocation logic, and spatio-temporal constraints of the IoT services are significant challenges. To solve these problems, a framework based on the event logic graph is proposed. It employs the event logic graph to describe the collaboration between IoT services and uses related methods to dynamically select appropriate services according to the spatio-temporal constraint. Specifically, event instances are first recognized from data generated by IoT services with a divide-and-conquer-based strategy. Next, event logic graphs are constructed based on the serial episode mining method. The related event prediction method is established for determining the service invocation logic. And a spatio-temporal constrained mechanism is designed to select the appropriate services. The proposed framework is applied to wheel manufacturing to promote autonomous production and is further compared with related works. The results show the feasibility of the proposed framework and its advantages in adaptability, dynamics, and spatio-temporal relevancy.
... A pre-eminent application domain for IoT is the smart home. A smart home is any regular home that is augmented with IoT devices [3]. The main objective of a smart home is to provide residents with efficiency and convenience [1]. ...
... For instance, a TV is located in a living room; a toaster service is located in a kitchen. We cluster services that are located in the same location (lines [2][3][4][5][6][7][8][9][10] in algorithm 1). When a TV service and a DVD service are located in a living room, service cluster of the living room is LS living = STV , SDV D . ...
Conference Paper
Full-text available
We propose a novel framework to detect conflicts among IoT services in a multi-resident smart home. A novel IoT conflict model is proposed considering the functional and non-functional properties of IoT services. We design a conflict ontology that formally represents different types of conflicts. A hybrid conflict detection algorithm is proposed by combining both knowledge-driven and data-driven approaches. Experimental results on real-world datasets show the efficiency of the proposed approach.
... Smart home is one of the preeminent application domains for IoT. A smart home is equipped with IoT devices that monitor the use of various "things" in the home to provide efficiency and convenience [2]. ...
... Therefore, each service usage habit is associated with a search space. We cluster service usage habits (SU H) based on location (lines [2][3][4][5][6][7][8][9][10] in algorithm 1). Services that are located in the same location form a cluster. ...
Conference Paper
Full-text available
We propose a novel conflict detection framework for IoT services in multi-resident smart homes. A fine-grained conflict model is developed considering the functional and non-functional properties of IoT services. The proposed conflict model is designed using the concept of entropy and information gain from information theory. We use a novel algorithm based on temporal proximity to detect conflicts. A set of experiments on real-world datasets are conducted to show the efficiency of the proposed approach.
... Smart home is one of the preeminent application domains for IoT. A smart home is equipped with IoT devices that monitor the use of various "things" in the home to provide efficiency and convenience [2]. ...
... Therefore, each service usage habit is associated with a search space. We cluster service usage habits (SU H) that are located in the same location (lines [2][3][4][5][6][7][8][9][10] in algorithm 1). When a TV service and a DVD service are located in a living room, SU H cluster of the living room is LSU H living = SU H T V , SU H DV D . ...
Preprint
Full-text available
We propose a novel framework to detect conflicts among IoT services in a multi-resident smart home. A fine-grained conflict model is proposed considering the functional and non-functional properties of IoT services. The proposed conflict model is designed using the concept of entropy and information gain from information theory. We use a novel algorithm based on temporal proximity to detect conflicts. Experimental results on real-world datasets show the efficiency of the proposed approach.
... After building the things graph, a Random Walk with Restart method is devised to find things relations. In our previous work [36], we propose an efficient algorithm PMiner to find spatio-temporal relationships among IoT services. A proximity metric is proposed to prune non-promising relationships among IoT services. ...
Preprint
Full-text available
We present a Cognitive Amplifier framework to augment things part of an IoT, with cognitive capabilities for the purpose of improving life convenience. Specifically, the Cognitive Amplifier consists of knowledge discovery and prediction components. The knowledge discovery component focuses on finding natural activity patterns considering their regularity, variations, and transitions in real life setting. The prediction component takes the discovered knowledge as the base for inferring what, when, and where the next activity will happen. Experimental results on real-life data validate the feasibility and applicability of the proposed approach.
Chapter
We propose a novel approach for predicting a resident’s future intentions in terms of what, how, when, and where he will do next. The intention model is learned from his previous interactions with various types of IoT services. In particular, we propose a multi-task learning approach for predicting resident’s future intention. The multi-task learning approach jointly learns the tasks of what, how, when, and where to boost the overall performance of the four tasks. We demonstrate the effectiveness and performance of our approach by conducting experiments on real-world datasets.
Conference Paper
Full-text available
We propose a novel conflict resolution framework for IoT services in multi-resident smart homes. An adaptive priority model is developed considering the residents’ contextual factors (e.g., age, illness, impairment). The proposed priority model is designed using the concept of the analytic hierarchy process. A set of experiments on real-world datasets are conducted to show the efficiency of the proposed approach.
Conference Paper
Full-text available
By bridging the physical and the virtual worlds, the Inter-net of Things (IoT) impacts a multitude of application domains, among which smart cities, smart factories, resource management, intelligent transportation, health and well-being to name a few. However, lever-aging the IoT within software applications raises tremendous challenges from the networking up to the application layers, in particular due to the ultra-large scale, the extreme heterogeneity and the dynamics of the IoT. This paper more specifically explores how the service-oriented architecture paradigm may be revisited to address challenges posed by the IoT for the development of distributed applications. Drawing from our past and ongoing work within the MiMove team at Inria Paris, the paper discusses the evolution of the supporting middleware solutions spanning the introduction of: probabilistic protocols to face scale, cross-paradigm interactions to face heterogeneity, and streaming-based interactions to support the inherent sensing functionality brought in by the IoT.
Article
Full-text available
We propose a model for clustering data with spatiotemporal intervals, which is a type of spatiotemporal data associated with a start-and an end-point. This model can be used to effectively evaluate clusters of spatiotemporal interval data, which signifies an event at a particular location that stretches over a period of time. Our work aims to deal with evaluating the results of clustering in multiple Euclidean spaces. This is different from traditional clustering that measure results in single Euclidean space. A new energy function is proposed that measures similarity and balance between clusters in spatial, temporal, and data dimensions. A large collection of parking data from a real CBD area is used as a case study. The proposed model is applied to existing traditional algorithms to solve spatiotemporal interval data clustering problem. Using the proposed energy function, the results of traditional clustering algorithms are compared and analysed.
Article
Full-text available
We propose a system, called EVolution of Long-term Composed Services (Ev-LCS), to address the change management issues in Long-term Composed Services (LCSs). A LCS is a dynamic collaboration between autonomous Web services that collectively provide a value-added service. It has a long term commitment to its users. We first present a formal model, which provides the grounding semantics to support the automation of change management. We then present a set of change operators that allow to specify a change in a precise and formal manner. We then propose a change enactment strategy that actually implements the changes. We develop a prototype system for the proposed Ev-LCS to demonstrate its effectiveness. We also conduct an experimental study to assess the performance of the change management approaches.
Conference Paper
Full-text available
Mining sequential patterns is an important issue for the complexity of temporal pattern discovering from sequences. Current mining approaches either require many times of database scanning or generate several intermediate databases. As databases may fit into the ever-increasing main memory, efficient memory-based discovery of sequential patterns will become possible. In this paper, we propose a memory indexing approach for fast sequential pattern mining, named MEMISP. During the whole process, MEMISP scans the sequence database only once for reading data sequences into memory. The findthen- index technique recursively finds the items which constitute a frequent sequence and constructs a compact index set which indicates the set of data sequences for further exploration. Through effective index advancing, fewer and shorter data sequences need to be processed in MEMISP as the discovered patterns getting longer. Moreover, the maximum size of total memory required, which is independent of minimum support threshold in MEMISP, can be estimated. The experiments indicates that MEMISP outperforms both GSP and PrefixSpan algorithms. MEMISP also has good linear scalability even with very low minimum support. When the database is too large to fit in memory in a batch, we partition the database, mine patterns in each partition, and validate the true patterns in the second pass of database scanning. Therefore, MEMISP may efficiently mine databases of any size, for any minimum support values.
Article
Full-text available
The increasing usage of smart embedded devices in business blurs the line between the virtual and real worlds. This creates new opportunities to build applications that better integrate real-time state of the physical world, and hence, provides enterprise services that are highly dynamic, more diverse, and efficient. Service-Oriented Architecture (SOA) approaches traditionally used to couple functionality of heavyweight corporate IT systems, are becoming applicable to embedded real-world devices, i.e., objects of the physical world that feature embedded processing and communication. In such infrastructures, composed of large numbers of networked, resource-limited devices, the discovery of services and on-demand provisioning of missing functionality is a significant challenge. We propose a process and a suitable system architecture that enables developers and business process designers to dynamically query, select, and use running instances of real-world services (i.e., services running on physical devices) or even deploy new ones on-demand, all in the context of composite, real-world business applications.
Article
We present a new two-level composition model for crowdsourced Sensor-Cloud services based on dynamic features such as spatio-temporal aspects. The proposed approach is defined based on a formal Sensor-Cloud service model that abstracts the functionality and non-functional aspects of sensor data on the cloud in terms of spatio-temporal features. A spatio-temporal indexing technique based on the 3D R-tree to enable fast identification of appropriate Sensor-Cloud services is proposed. A novel quality model is introduced that considers dynamic features of sensors to select and compose Sensor-Cloud services. The quality model defines Coverage as a Service which is formulated as a composition of crowdsourced SensorCloud services. We present two new QoS-aware spatio-temporal composition algorithms to select the optimal composition plan. Experimental results validate the performance of the proposed algorithms.
Conference Paper
In this work, a system for recognizing activities in the home setting using a set of small and simple state-change sensors is introduced. The sensors are designed to be “tape on and forget” devices that can be quickly and ubiquitously installed in home environments. The proposed sensing system presents an alternative to sensors that are sometimes perceived as invasive, such as cameras and microphones. Unlike prior work, the system has been deployed in multiple residential environments with non-researcher occupants. Preliminary results on a small dataset show that it is possible to recognize activities of interest to medical professionals such as toileting, bathing, and grooming with detection accuracies ranging from 25% to 89% depending on the evaluation criteria used.
Book
This is the third edition of the premier professional reference on the subject of data mining, expanding and updating the previous market leading edition. This was the first (and is still the best and most popular) of its kind. Combines sound theory with truly practical applications to prepare students for real-world challenges in data mining. Like the first and second editions, Data Mining: Concepts and Techniques, 3rd Edition equips professionals with a sound understanding of data mining principles and teaches proven methods for knowledge discovery in large corporate databases. The first and second editions also established itself as the market leader for courses in data mining, data analytics, and knowledge discovery. Revisions incorporate input from instructors, changes in the field, and new and important topics such as data warehouse and data cube technology, mining stream data, mining social networks, and mining spatial, multimedia and other complex data. This book begins with a conceptual introduction followed by a comprehensive and state-of-the-art coverage of concepts and techniques. Each chapter is a stand-alone guide to a critical topic, presenting proven algorithms and sound implementations ready to be used directly or with strategic modification against live data. Wherever possible, the authors raise and answer questions of utility, feasibility, optimization, and scalability. relational data. -- A comprehensive, practical look at the concepts and techniques you need to get the most out of real business data. -- Updates that incorporate input from readers, changes in the field, and more material on statistics and machine learning, -- Scores of algorithms and implementation examples, all in easily understood pseudo-code and suitable for use in real-world, large-scale data mining projects. -- Complete classroom support for instructors as well as bonus content available at the companion website. A comprehensive and practical look at the concepts and techniques you need in the area of data mining and knowledge discovery.