KONERU LAKSHMAIAH COLLEGE OF ENGINEERING 

(Approved By AICTE, Accredited By (NBA & ISO 9001 2000 Certified) 
GREEN FIELDS, VADDESWARAM.

        Term Paper Report on 
     
     A data mining approach for location 
prediction in mobile environments 
 
 
 
 
Guided by                 Submitted By 
P.V.V.Prasad                  kvr.Anjaneyulu (L8CS340) 
Lecturer                  K.Venkatesh      (Y7CS312) 
Department of CSE              N.Mahesh           (Y7CS244) 
                                                       P.Sangram         (Y7CS288)

 
 
 
DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING     
2010­2011 
 Literature Survey
1. Predictive mobility support for QoS provisioning in mobile wireless
environments

A. Aljadhai, T. Znati

This paper presents a predictive and adaptive scheme to support timed-QoS that
guarantees in pico- and micro-cellular environments.

The proposed scheme integrates the mobility model into the service model to achieve
efficient network.

The mobility model uses a probabilistic approach to determine the most likely cluster
to be visited by the mobile unit.

The admission control is invoked when a new call arrives or an existing call performs
a handoff to verify the feasibility of supporting the call.

2. Effective Prediction of Web-user Accesses:A Data Mining Approach.

A. Nanopoulos, D. Katsaros, Y. Manolopoulos.

In this paper, the focus is on web-prefetching, because of its importance in reducing

user perceived latency present in every Web-based application.

Prefetching refers to the mechanism of deducing the forthcoming page accesses of a

client, based on access log information

The most successful approach towards this goal has been the exploitation of the
user’s access history to derive predictions.

The caching of web documents at various points in the network (client, proxy, server)
has been developed to reduce the latency.
3. Movement-based location update and selective paging for PCS networks

I.F. Akyildiz, S.M. Ho, Y.-B. Lin

This paper introduces a mobility tracking mechanism that combines a movement-

based location update policy with a selective paging scheme.

In fact, each mobile terminal only keeps a counter of the number of cells visited.

ARCHITECTURE OF PCS NETWORKS

We propose a selective paging scheme which significantly decreases

the location tracking cost under a small increase in the allowable paging delay.

We introduce an analytical model for the proposed location tracking mechanism

which captures the mobility and the incoming call arrival patterns of each mobile
terminal.
4. Predictive distance-based mobility management for PCS networks.
B. Liang, Z. Haas.

This paper presents a mobile tracking scheme that exploits the predictability of user
mobility patterns in wireless PCS networks

Based on the Gauss-Markov model, a mobile’s future location is predicted by the

network based on the information gathered from the mobile’s last report of location
and velocity.

A mobile shares the same prediction information with the network and reports its new
location whenever it reaches some threshold distance away from the predicted
location.

We describe an analytical framework to evaluate the cost of mobility management for

the proposed predictive distance-based scheme.

5. Clustering Mobile Trajectories for Resource Allocation in Mobile

Environments .
D. Katsaros, A. Nanopoulos, M. Karakaya, G. Yavas,
O. Ulusoy, Y. Manolopoulos.

In this paper we deal with the problem of how to predict the position of a mobile
client.

We propose a new algorithm to discover user mobility patterns from collections of

recorded mobile trajectories and use them for the prediction of movements and
dynamic allocation of resources.

The network mobility gives rise to some new and important problems like location
management and the bandwidth allocation problems.

The location management allows the system to keep the user’s location knowledge
(exact or approximate) in order to be able to locate him.

The paging procedure consists of sending paging messages in all the locations (cells)
where the mobile client could be located.
6. GPS-based predictive resource allocation in cellural networks.
S. Rajagopal, R.B. Srinivasan, R.B. Narayan, X.B.C. Petit.

Here we propose that GPS receivers be interfaced with the mobile terminals to
provide the location information.

A new architecture has been proposed, in which the cell assignment is handled
dynamically through estimation of the location of mobile terminals by a resource
allocation center.

The simulation results show that the GPS based prediction is able to reduce the hand-
off call dropping by as much as 38% for heavy traffic conditions.

The call dropping rate can be minimized by allocating resources in neighbouring cells
based on the location of the mobile devices.

7. Two user location strategies for personal communications services.

S. Mohan, R. Jain.

It is necessary to locate users who move from place to place.

The strategies commonly proposed are two-level hierarchical strategies, which

maintain a system of home and visited databases-home location register (HLR) and
visitor location register (VLR)-to keep track of user locations.

The standards exist for carrying out two-level hierarchical strategies using HLRs and
VLRs (EIA/TIA) and Interim Standard 41 (IS-41)

The authors introduce these two strategies for locating users. Different forms of
mobility in the context of PCS are discussed.
8. Exploiting data mining techniques for broadcasting data in mobile
computing environments.
Y. Saygin, O. Ulusoy.

We propose two methods aiming to reduce client access latency of broadcast data.

Our methods are based on analyzing the broadcast history using data mining
techniques.

With the first method, the data items in the broadcast disk are organized in such a
way that that the second method focuses on improving the cache hit ratio to be able to
decrease the access latency.

It enables clients to prefetch the data from the broadcast disk based on the rules
extracted from previous data request patterns.

The proposed methods are implemented on a Web log to estimate their effectiveness

PROBLEM STATEMENT

• In our work, we collect the movement trajectories of a user in the form of T =

h[(id1, t1), (id2, t2), . . ., (idk, tk)]

• If the mobile user stays in a cell idi more than a threshold value, before

moving to another one idi+1 at ti+1, we assume that his trajectory up until now

id1, . . ., idi ends here, and at idi+1 a new trajectory is started.

• We name the trajectories obtained by the above procedure as user actual paths
(UAPs).

• We assume that we have UAPs which have the form U = h(c1, c2, . . ., cn)

• We name the frequently followed trajectories as user mobility patterns (UMPs).

• Mining of the UMPs enables us to generate mobility rules.

• By considering the mobility rules and the trajectory of a user, we predict the

next inter-cell movement of the user.

 EXISTING SOLUTIONS

The Mobile Motion Prediction (MMP) Algorithm

• users moving behavior is modeled as repetitions of some elementary

movement patterns which are indeed circular and straight line patterns.

• In order to estimate the future location of a user, a mobile motion prediction

(MMP) algorithm is proposed.

• However, the MMP algorithm is highly sensitive to random movements of the

user. It is reported that as the random movements of the user increase the

performance of MMP decreases linearly.

Sequential pattern mining

• sequential pattern mining is applied to the domain of predictive Web

prefetching.

• Webprefetching can be defined as deriving users future requests for Web

documents based on their previous requests.

• For effectively predicting the users_ future requests, user access patterns are

mined from the web logs of users_ previous requests and then these patterns

are used for prefetching.

 Web prefetching Architecture

Dynamic clustering based prediction

• DCP is used for discovering user mobility patterns from collections of recorded

mobile trajectories, and then these patterns are used for the prediction of

movements and dynamic allocation of resources.

• At each iteration of the clustering algorithm, the two most similar clusters are

merged to form a new cluster.

• After each merge operation, the representatives of new cluster are found to be

the union of representative sets of the merged clusters.

• The merge operation continues until the number of the clusters is reduced to a

predefined value.

• A mobile users next trajectory is predicted by finding the best matching

representative with its current trajectory.