Context Based Content Aggregation

for Social Life Networks

Maneesh Mathai and Athula Ginige

School of Computing, Engineering And Mathematics, University of Western Sydney, New South Wales, Australia

Keywords:

Context Modelling, Content Aggregation, Social Life Networks, Mobile Application.

Abstract:

It is extremely useful to have right information at the right time. Social Life Networks (SLN) extend the

capabilities of current social networks by combining them with the technological advances now found in

Smartphones that include myriad of sensors and multimedia input and output capabilities to provide essential

information to support livelihood activities. The challenge is to provide this information within the required

context. For this we need to model the context by acquiring the physical data to provide meaningful abstrac-

tions with respect to the application domain and the needs of the users. We have developed a physical context

model based on user proﬁle, location, time and activity and a mapping to match the logical context of vari-

ous data sources from which we can get the required information. Based on this model we have developed a

SLN for farmers in Sri Lanka to provide agricultural information in the context of farming life cycle stages,

location of their farm land, cultivation season and other economic parameters. In the ﬁeld trails there was

unanimous agreement among farmers that this application is very useful for them because they were able to

get the required information in context.

1 INTRODUCTION

Due to lack of access to current information, often

farmers tend to grow the same crop within a region,

and this is causing potential over supply of crops

(Hettiarachchi, 2011). Farmers come to know, or re-

alize, there is an oversupply only when they bring

their harvest to the market, and the oversupply re-

duces market price for the crop, disadvantaging the

farmers. This is a frequent scenario happening in de-

veloping countries (Hettiarachchi, 2011). Neither the

farmers nor government agencies are able to make

necessary adjustments for lack of timely information

on what farmers plan to cultivate, or have cultivated.

The yield could also be affected by various factors in-

cluding availability of water, weather, and pests.

(De Silva et al., 2012) have shown that farmers

can make an informed decision on what crop to grow

if they have access to a mobile phone based informa-

tion system to inquire what others in that region are

growing. The information system can provide this in-

formation only if most farmers use this system and

indicate what crop they plan to grow. Aggregating

the information provided by the farmers the informa-

tion system can also inform the government agen-

cies monitoring and managing agriculture sector in

a country, fertilizer and pesticide suppliers and po-

tential buyers what has already been grown for better

management of the overall crop production.

Mobile phone usage in the world has grown

rapidly including among people in developing coun-

tries. At present, 90% of the world population is cov-

ered by a mobile signal, 128% of the world population

has a mobile subscription and in developing coun-

tries the subscription rate is 89% (ict, 2013). Further

Smartphone prices are rapidly decreasing and now are

comparable to a cost of a basic mobile phone few

years ago. A Smartphone can be considered as a

sensor in the hands of a human capable of capturing

user input as text, voice or gestures and other envi-

ronmental parameters using build in sensors such as

GPS, camera etc. It is also capable of communicating

with the user using range of media types; text, images,

video, and audio.

An International Collaborative research project

was started to develop a Social Life Network (SLN)

(ict, 2013) ; a mobile based information system to

support livelihood activities of people in developing

countries. Social Life Networks (SLN) tries to ex-

tend the capabilities of current social networks by

combining them with the technological advances now

found in Smartphones that include myriad of sensors

570

Mathai M. and Ginige A..

Context Based Content Aggregation for Social Life Networks.

DOI: 10.5220/0004596205700577

In Proceedings of the 8th International Joint Conference on Software Technologies (ICSOFT-PT-2013), pages 570-577

ISBN: 978-989-8565-68-6

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

and multimedia input and output capabilities (Jain and

Sonnen, 2011). An essential component of the SLN

architecture is a module that provides information to

users in context. In order to get a deeper insight into

research challenges and to investigate possible solu-

tions we selected a speciﬁc real world problem to

work on. The ﬁrst SLN was developed for farmers

in Sri Lanka to address the over production problem

mentioned above.

2 BACKGROUND

2.1 Components for SLN

(Jain and Sonnen, 2011; Ginige; et al., 2012) pro-

posed that there are a few basic components for real-

izing the vision of social life networks. Data coming

from multiple users and heterogeneous devices needs

to be wrapped into a common format and made ac-

cessible to the system. Logically the data needs to

be translated from localized sensor/human input to

higher level situational abstractions. There is also an

encompassing issue of user engagement. Both intrin-

sic and extrinsic factors matter, but enhanced feed-

back and user motivation are key aspects of it.

The biggest catalyst for the adoption of the tradi-

tional Web was the presence of search engines which

routed users to their desired resources (static web

pages). A situation analysis performs a similar role

in the dynamic social life networks i.e. routing the

users to the appropriate resources based on situation

detected. In this paper we propose that to carry out

the situation analysis, it is vital to identify the context

and using the contextual information will result in a

more personalized set of results for the user.

2.2 Context

During the past two decades, researchers have devel-

oped techniques that enable systems to adapt to their

users in many different ways (Seher et al., 2007). One

of the major research directions for humancomputer

interaction (HCI) and Information Retrieval (IR) has

been exploring the novel forms of interaction that can

be achieved by integrating computer technology with

the everyday physical world in which we live and

work. Ubiquitous or pervasive computing represents

a powerful shift in computation, where people live,

work and play in a seamless computer-enabled envi-

ronment and people are surrounded by computing de-

vices and a computing infrastructure that supports us

in everything we do (Poslad, 2011; ?; ?). Ability to

accurately represent user context is very important to

make optimum use of these smart environments. For

this we need to model the context by acquiring the

physical data to provide meaningful abstractions with

respect to the application domain and the needs of the

users interacting with the application.

A qualiﬁed deﬁnition of context is given by (Dey,

2001). In this work the term context is deﬁned as fol-

lows:

Context as any information that can be used to char-

acterize the situation of an entity. An entity is a per-

son, place, or object that is considered relevant to

the interaction between a user and an application, in-

cluding the user and applications themselves.

After deﬁning context, context modelling is used

to handle context information for the different entities

of the application.

2.2.1 Need for Context

Human communications are quite successful and

seem to be very easy. This is mainly due to the rich-

ness of the language used, and the common under-

standing humans have about how the world works

(Dey, 2001). Human communication is more success-

ful when there is an implicit understanding of the ev-

eryday situations of others who take part in the com-

munication.

For example, in a conversation between two

friends, the question How was the game? might not

require any further elaboration. This question is in-

complete, as it does not explicitly describe the game.

Still, the respondent can understand the question.

This is possible if the humans involved in the commu-

nication have a shared context which includes know-

ing the everyday situations of each others lives. More-

over, the type of the answer expected is not mentioned

in the question. However, it is understood that the an-

swer is not only about who won the game, but also

should include descriptions of the game, such as the

scores.

For an automated system to have the ability of

bringing this contextual information in the interaction

between the human and the computer is challenging.

By improving the computers access to the context, the

richness of communication in humancomputer inter-

action can be increased, giving more useful compu-

tational services. Furthermore, personalized informa-

tion services help to improve the conversational band-

width in a humancomputer interaction.

2.2.2 Context-speciﬁc information for

Agriculture

Information must be relevant and meaningful to farm-

ers, in addition to being packaged and delivered in a

ContextBasedContentAggregationforSocialLifeNetworks

571

way preferred by them (Diekmann et al., 2009) cited

in (Babu et al., 2012). Context-speciﬁc information

could have a greater impact on the adoption of tech-

nologies and increase farm productivity for marginal

and small agricultural landholders (Samaddar, 2006)

cited in (Babu et al., 2012). However, making infor-

mation context-speciﬁc is more resource intensive. It

requires information at the farm level, which could

vary spatially and temporally, and with different de-

grees of speciﬁcity (Garforth et al., 2003) cited in

(Babu et al., 2012). Despite the additional cost and

time associated with generating localized content, this

content could be more relevant and useful in meeting

farmers information needs (Cecchini and Scott, 2003)

cited in (Babu et al., 2012).

(Babu et al., 2012) discuss clearly the importance

of contextualized information and knowledge for the

farmers in India. They further explain how effective

this knowledge can improve their productivity and in-

come since this information is more relevant to their

farm enterprises and better reﬂects needs of the farm-

ers. They therefore recommend that the existence of

context-speciﬁc and relevant information should be

considered when developing approaches for farmers.

In this paper, we describe context speciﬁc to the

farmers in Sri Lanka and the approach we developed

to design the context expansion module to provide

context-speciﬁc information and knowledge to farm-

ers. It further discusses the design and implementa-

tion of the ﬁrst version of a mobile application for

framers. The remainder of the paper is organized as

follows. Section 3 presents the methodology for mod-

elling context. Section 4 describes the high level ar-

chitecture and implementation of the mobile applica-

tion. Finally, section 6 concludes the paper.

3 MODELING CONTEXT

Every activity that a user can perform can be subdi-

vided into sub tasks or phases. User needs speciﬁc

information based on each phase.

In farming domain, (De Silva et al., 2012) has

identiﬁed that farmers in Sri-Lanka need speciﬁc in-

formation rather than generic information. For in-

stance, farmers need agricultural information relevant

to their situation such as the location of their farm

land, their economic condition, their interest and be-

lief, need and available equipments etc.

(De Silva et al., 2012) carried out a causal analysis

to determine the factors that inﬂuence farmers deci-

sion making at various stages of the farming lifecycle

and in that process they identiﬁed what speciﬁc in-

formation is required in each stage. The causal anal-

ysis was carried out through a series of surveys. In

this process, they also looked at the various informa-

tion sources currently available for farmers following

which (De Silva et al., 2012) determined how the in-

formation needs to ﬂow to the farmers. (De Silva

et al., 2012) postulated that the information should be

made available to farmers in context to make mean-

ingful use of the information and that the information

needs of farmers change with each stage of the farm-

ing cycle.

Deeper information need analysis revealed that

the farmers need two types of information for each

stage; dynamic information such as current extent of

crop cultivation, market prices etc. and more stable

static information such as crop types, cultivars, suit-

able pesticides, fertilizer, previous market prices etc

(De Silva et al., 2012; Walisadeera et al., 2013). The

agriculture domain knowledge obtained from (Wal-

isadeera et al., 2013), also suggests that farm envi-

ronment, types of farmers, farmers preferences and

farming stages are the important factors that needs to

be considered for ﬁltering relevant information.

3.1 Physical and Logical Context

When developing context aware applications we need

to consider two types of contexts. The physical

context attributes are the raw environmental param-

eters which are captured in real time using sensors

or pre stored in the system while the logical context

attributes are concepts associated to physical data,

which provide meaningful abstractions with respect

to the application domain and the needs of the users

interacting with the application.

Each user of the application will have different

physical values according to their settings. Every ap-

plication would have its own logical interpretation of

the stored physical data. The system needs to derive

the logical application context from the physical data

to get information in context.

3.2 Modelling Context in Farming

Domain

To develop a context model for the farming domain

we had to ﬁrst identify the relevant physical context,

the logical context and the mapping between the two.

For example one of the attribute of the physical con-

text that is stored by the system is the geo coordinates.

The geo-coordinates can be interpreted in many pos-

sible ways like agro ecological zones or administra-

tive districts. If the goal of the application is to deter-

mine the environmental attributes of the region then

the application can map the geo-coordinates into agro

ICSOFT2013-8thInternationalJointConferenceonSoftwareTechnologies

572

ecological zones otherwise if the goal is to obtain the

market price of the region, then the application needs

to determine under which administrative district the

given geo-coordinates belong and then query the ap-

propriate information source to get the relevant data.

Figure 1: Physical to Logical Context Mapping.

Thus, it can be seen that the same physical context

attribute can be mapped into different logical context

attribute according to the information requirements of

each of the application. So, one of the important ca-

pabilities of the context expanding application would

be to facilitate this conversion of raw physical data to

match the required logical context. Different appli-

cation can have space- time- user related information

and activities represented different ways which would

allow different application to identify and querying

different databases static or dynamic, based on the

requirements.

3.3 Designing the Generic Model for

Physical Context

Physical context can be modelled using user proﬁle,

space, time and activity as major sub categories. Each

sub category will consist of further sub categories and

attributes.

Proﬁle Modelling: This consists of the user con-

text, such as the users proﬁle and preferences. Some

of these are general or domain-independent to a user.

Examples include the users personal characteristics

such as profession, age and gender. Some other user

preferences are speciﬁc to the domain or application

in consideration. These could be the users level of

knowledge of particular topics, and the users level of

interest in particular topics. Starting from the identity

of a user, it is possible to obtain individual context

information. This can be achieved by having a regis-

tration system which manages the basic user proﬁle.

Spatial Information: This context type describes

aspects relating to the spatial extent of the user con-

text. It can contain attributes like location, direc-

tion and speed. One of the most common contexts

used is the location of the user. Location-awareness

is the most important part of context-awareness for

mobile computing systems. The spatial region of

interest deﬁnes the physical boundaries and may be

of any shape. Spatial context is represented as re-

gions, spatial relationships and geometric coordi-

nates. In a regional representation, a boundary is

speciﬁed as regions, such as Australia, Sydney or Par-

ramatta. These representations can be hierarchically

organized. Spatial relationships can be direction rela-

tionships (above, below, or north of, southwest of),

topological relationships (near, far, around, within,

adjacent, inside), and metric relationships (distance).

The geometric coordinates specify points or areas in a

metric space that represent the latitude, longitude and

elevation above sea level.

A users location can be sensed using Active

Badges, radar, video cameras or GPS (Global Po-

sitioning System) units. GPS sensors have become

available in very small package sizes, enabling their

integration in mobile devices. The level of precision

required by the GPS module will depend on the in-

formation requirement of the application. For some

application there could be instances where multiple

geo-coordinates may be made associated with an in-

dividual user proﬁle.

Figure 2: Generic Model for Physical Context.

Temporal Information: This context type de-

scribes aspects relating to time. Time is a fundamen-

tal variable, since the context is dynamic and changes

with time. A temporal context can be represented as

time instants or absolute time references, time inter-

vals, periodic descriptors, and temporal relationship.

A time instant or absolute time reference is an in-

stant of time that is an absolute moment of time. It

can be different in time units such as 9.00 a.m. and

11/10/2005. A time interval is based on calendar units

such as year, today, morning. It is the time interval

between two absolute time references.

ContextBasedContentAggregationforSocialLifeNetworks

573

Activity Information: The task context can be de-

scribed with explicit goals, tasks, actions, activities,

or events. The task context could be speciﬁc to the

domain in which those tasks are performed. It can be

decomposed into sub-tasks.

4 SOCIAL LIFE NETWORK

APPLICATION FOR FARMERS

(Ginige and Ginige, 2011) states that connectivity

and empowerment as the two important characteris-

tics that needs to be addressed when developing an

application for farmers. Now let us look at an exem-

plary scenario to show how an application can work

in an agriculture domain.

Example: A good example to describe the advan-

tage of using a context aware application would be

a situation in which the farmer decide what crop to

grow?. The system identiﬁes that the farmer is in the

ﬁrst stage of farming and thus only needs to concern

itself with the subset of information that the farmer

might need at this stage. This sub set of information

needs has already been identiﬁed through the analysis

of the domain (Lokanathan and Kapugama, 2012) as

shown in table 1.

Table 1: Information needs of Farmers.

Information needs of Farmers

in Crop Selection Stage.

Current market prices for a speciﬁc crop(s)

in the speciﬁc market that I sell at

Current market prices for a speciﬁc crop(s)

in the speciﬁc market that I sell at

Current market prices for a speciﬁc crop(s)

in market(s) other than what I sell at

Expected future market prices for a speciﬁc crop(s)

around the time when your crops will be ready for

harvesting

Information on ﬁnance

(formal and informal sources, the cost involved etc)

Information on govt. schemes

(including subsidies and minimum support prices)

and policies on agriculture (current as well as changes)

Information on higher yield crops

Information on best farming practices

including how to grow a particular crop

Information on crop diseases and how to solve them

Information on input availability and associated costs

Information on labour availability and associated costs

Information on land availability and associated costs

Information on farming machinery/equipment

and associated costs

Information on electricity timings

Information on water availability

Information on weather

The system then provides a list of crops suitable

to the farm of the farmer based on the fact that the

farm is located in a particular agro ecological zone of

Sri-Lanka and therefore a matching can be done to

produce a list of crops that could be grown in that

ecological zone. This list is further limited by the

season in which the farmer intends to grow the crop.

This list of crop can then be ranked according to the

current production level if that dynamic information

is available to the system. To empower the farmers

each crop would be associated with vital information,

which would then assist the farmers in their decision

making process.

In this particular scenario, the system is able to

ﬁlter, rank and provide information about different

crops by identifying the context of the farmer. The

context of the farmer is dictated by the ecological

zone in which the farm is located, the farming stage,

the season in which the farmer intends to grow the

crop, and the personal preferences of the farmer.

4.1 Components of SLN4Farmers

An appropriate infrastructure for social life networks

should support most of the tasks required to identify

the context. The context expansion model needs to

focus on a scalable context management while a data

manger should be responsible for reasoning and con-

cept representation. For example if the farming stage

is crop selection then the context expansion module

creates the query what crop to grow?.

Figure 3: High Level Architecture for SLN.

Conceptually the data manager can be grouped

into two, ontology based knowledge repository for

static information and real time aggregation module

for dynamic information. (Walisadeera et al., 2013)

has shown how ontology can be used to ﬁnd a re-

sponse to queries within a speciﬁed context in the

domain of agriculture. This structured view is essen-

tial to facilitate knowledge sharing, knowledge aggre-

gation, information retrieval, and question answering

while the real time aggregation module ingests and

ICSOFT2013-8thInternationalJointConferenceonSoftwareTechnologies

574

manages the sensory and dynamic data, based on the

attribute values available at run time.

The focus of this paper, the context expansion

module has the application data that is speciﬁc to the

application domains characteristics which is used to

carry out the logical interpretations and a physical

context model that is responsible for keeping the con-

text state updated for each individual user of the ap-

plication. The context characterizing properties that

change with spatial and temporal attributes is cap-

tured by the physical context model.

The context expansion module provide the follow-

ing functionality :

• The deﬁnition and structure of context model that

is used to represent the task being carried out by

the speciﬁc application.

• The acquisition of data at run time that is used by

the context model to provide personalized infor-

mation in context.

4.2 Context Expansion

The context expansion module identiﬁes information

need of the sub task of the activity performed by the

user. This information along with the logical interpre-

tation of the application of the stored physical context

of the user is used to ﬁlter and provide the required

information.

Figure 4: Conceptual Model for Context Expansion.

Let us look at one of the working example of the

application. The farmer logs into the system, the sys-

tem asks the farmer to register his farm. This is gov-

erned by the basic application domain knowledge that

a farmer might have multiple farms associated with

him. The context model allows us to capture this in-

formation through the user proﬁle. The current stage

of farming is also captured in the user proﬁle and

having the relationship that each farmer has multiple

farms allows us to capture the farming stage of each

farm associated with the farmer. The GPS system of

the mobile is used to capture the geo-coordinates for

each farm.

Figure 5: Input Output Table for Context.

Once the farmer selects a farm the information

speciﬁc to that farm will be made available by the

system. If the farmer selects crop planning the con-

text expansion module checks the farming stage of

that particular farm and then provide information rel-

evant to that farm for that particular farming stage by

creating a corresponding query.

For example if the farming stage is crop selection

then the context expansion module creates the query

what crop to grow?.

This query is further expanded by different blocks

of context expansion module as shown in the ﬁgure

3. Thus the query createTd by the context expansion

module is expanded from What crop to grow? into

What crop to grow in Yala with values agro ecolog-

ical zone=DL1, Low country, Dry zone; Rainfall =

100, Elevation = Low. This query is then passed on to

the data manager where agriculture ontology is used

to obtain a list of crops. For each of the individual

crop, a query is generated and sent to the aggregation

module, to obtain the current production level of each

vegetable.

4.3 Implementation

To test the concept an end to end prototype was devel-

oped. Focus was on user interface, usability and ac-

ceptance of the system by the end-users. The design

process was based on the research done by (De Silva

et al., 2012) and took into account several factors such

as users level of literacy, familiarity in using the de-

vice, users cultural background, and language beliefs.

ContextBasedContentAggregationforSocialLifeNetworks

575

Scenarios based on (Di Giovanni et al., 2012) studies

were used to identify a list of non-functional require-

ments. The above architecture was implemented as a

mobile up for Android phoned connected to a back-

end server through web services.

The application model speciﬁes the following in-

formation:

• The user is a farmer and that he can have multiple

farms and multiple mobiles.

• Each farm has geo-coordinates, season and the

farming stage associated with it.

• There is a set of information required for each

farming stage.

Figure 6: Login and Farm Selection.

Figure 7: Crop selection and List of crops.

Figure 6 shows the login screen. It has been de-

signed to handle the situation of farmers having mul-

tiple mobiles and also for the authentication process.

The application domain dictates the fact that a farmer

might have multiple farms and since each farm has

its own unique characteristics, the farmer is asked to

select a unique farm at the start of the application.

Figure 7 represents the higher level of grouping of

information needs in the agriculture domain, which

Figure 8: Crop Characteristics.

caters to the different information needs at different

stages of farming. In the ﬁrst version, the crop plan-

ning stage was implemented, so when the farmer se-

lects this button, the system provides the user with a

list of suitable crops as shown in ﬁgure 8. The farmer

can ﬁnd more information about a particular crop by

selecting the individual crop. The farmer can on this

screen indicate which crop he plans to grow and in

what quantity. This information is stored as part of

the user proﬁle and is associated with the farm of the

farmer.

5 RESULT

The usability ﬁeld test with 32 farmers in Sri Lanka

was done to prove the effectiveness of the initial solu-

tion and positive response was obtained.

There was universal agreement among the farm-

ers participated in the ﬁeld trial to varying degree

(strongly agree, agree, moderately agree) to the ques-

tion All information for the crop choosing stage is

provided. This information was provided to the farm-

ers using farmer context via low cost Smart phones

running Android operating system.

6 CONCLUSIONS

To understand the research challenges and to derive

possible solutions to provide context based informa-

tion in Social Life Networks, we have taken a con-

crete example in the form of an application for farm-

ers that could meet the information needs within the

farmers context. To represent information in context,

we have developed an approach to model users phys-

ical context and carried out the mapping to obtain

ICSOFT2013-8thInternationalJointConferenceonSoftwareTechnologies

576

the logical context required by various data managers

based on the current attribute values of the physical

context. The context expanding module of the appli-

cation was responsible for facilitating the conversion

of raw physical data to match the required logical con-

text of the application.

The solution described in this paper have been

tested by creating a mobile application which has al-

lowed us to prove that the solution is feasible and

meets the information needs of farmers in Sri Lanka.

In the future we plan to improve the context expansion

module by inferring the intent of the users from their

interactions with the system (Deufemia et al., 2013).

The current application is a speciﬁc instance of the

SLN project and we plan to create a generalized ar-

chitecture that would be useful in creating many ap-

plication for SLN.

REFERENCES

(2013). The world in 2013: Ict facts and ﬁgures.

Babu, S. C., Glendenning, C. J., Okyere, K. A., and Govin-

darajan, S. K. (2012). Farmers’ information needs and

search behaviors: Case study in tamil nadu, india. In

2012 Conference, August 18-24, 2012, Foz do Iguacu,

Brazil, number 126226. International Association of

Agricultural Economists.

Cecchini, S. and Scott, C. (2003). Can information and

communications technology applications contribute to

poverty reduction? lessons from rural india. Informa-

tion Technology for Development, 10(2):73–84.

De Silva, L. N., Goonetillake, J. S., and Wikramanayake,

G. N. (2012). Towards using ict to enhance ﬂow of

information to aid farmer sustainability in sri lanka.

In Proceedings of the 23rd Australasian Conference

on Information Systems 2012, pages 1–10. ACIS.

Deufemia, V., Giordano, M., Polese, G., and Simonetti,

L. M. (2013). Exploiting interaction features in user

intent understanding. In Web Technologies and Appli-

cations, pages 506–517. Springer.

Dey, A. K. (2001). Understanding and using context. Per-

sonal and ubiquitous computing, 5(1):4–7.

Di Giovanni, P., Romano, M., Sebillo, M., Tortora, G., Vi-

tiello, G., Ginige, T., De Silva, L., Goonethilaka, J.,

Wikramanayake, G., and Ginige, A. (2012). User cen-

tered scenario based approach for developing mobile

interfaces for social life networks. In Usability and

Accessibility Focused Requirements Engineering (Us-

ARE), 2012 First International Workshop on, pages

18–24. IEEE.

Diekmann, F., Loibl, C., and Batte, M. T. (2009). The

economics of agricultural information: factors af-

fecting commercial farmers information strategies in

ohio. Applied Economic Perspectives and Policy,

31(4):853–872.

Garforth, C., Angell, B., Archer, J., and Green, K. (2003).

Improving farmers’ access to advice on land manage-

ment: lessons from case studies in developed coun-

tries. Overseas Development Institute. Agricultural

Research Extension Network (AGREN).

Ginige;, A., Ginige;, T., and Richards, D. (2012). Architec-

ture for social life network to empower people at the

middle of the pyramid,.

Ginige, T. and Ginige, A. (2011). Towards next generation

mobile applications for mops: investigating emerging

patterns to derive future requirements. In Advances in

ICT for Emerging Regions (ICTer), 2011 International

Conference on, pages 117–118. IEEE.

Hettiarachchi, S. (2011). Leeks cultivators desperate as

price drops to record low. Sunday Times. Sri Lanka.

Jain, R. and Sonnen, D. (2011). Social life networks. IT

Professional, 13(5):8–11.

Lokanathan, S. and Kapugama, N. (2012). Smallhold-

ers and micro-enterprises in agriculture: Informa-

tion needs and communication patterns. LIRNE asia,

Colombo, Sri Lanka, pages 1–48.

Poslad, S. (2011). Ubiquitous computing: smart devices,

environments and interactions. Wiley.

Samaddar, A. (2006). Traditional and posttraditional: A

study of agricultural rituals in relation to technological

complexity among rice producers in two zones of west

bengal, india. Culture & Agriculture, 28(2):108–121.

Seher, I., Ginige, A., and Shahrestani, S. (2007). A per-

sonalized query expansion approach using context. In

Intelligent Environments, 2007. IE 07. 3rd IET Inter-

national Conference on, pages 383–390. IET.

Walisadeera, A. I., Wikramanayake, G. N., and Ginige1, A.

(2013). An ontological approach to meet information

needs of farmers in sri lanka.

ContextBasedContentAggregationforSocialLifeNetworks

577