Automatic Updating of Computer Games Data Warehouse for Cognition

Identiﬁcation

∗

C. I. Ezeife

, Rob Whent

, Dragana Martinovic

, Richard Frost

, Yanal Alahmad

and Tamanna Mumu

School of Computer Science, University of Windsor, Windsor, Ontario, N9B 3P4, Canada

OTEP Inc, 13300 Tecumseh Road East, Suite 366, ON N8N 4R8, Tecumseh, Ontario, Canada

Faculty of Education, University of Windsor, Windsor, Ontario, N9B 3P4, Canada

Keywords:

Cognitive Skill Mapping, Computer Games, Automatic Database Integration, Automatic Database Mining.

Abstract:

This paper describes the algorithms (called OTEP DW auto) for automatically updating the integrated games

data warehouse and cognitive proﬁle data sources for purposes of identifying child’s cognitive skill level. The

techniques described in this paper represent an extension to the data integration engine adopted by an online

product called “Thriver” developed by OTEP Inc. (Online Training & Evaluation Portal). OTEP focuses on

using the Internet, natural playing environment for online computer games to give parents and care-givers

automated opportunity to screen and follow their children’s cognitive development. Current data integration

efforts of the system when new games (such as speech games) are added or new cognitive skills matrix are

added would require manual re-coding of the system which is a costly and time-consuming process. The

cognitive skills matrix maps cognitive skills level of games player such as “basic reading level is good” to

their games performance in comparison to the norms of other players. The proposed OTEP DW auto is

capable of building the OTEP data warehouse schema automatically, thus seamlessly extracting, cleaning and

propagating data from various data sources. It also provides a dynamic GUI-based interface for answering

tens of frequently asked cognition-related questions.

1 INTRODUCTION

With the proliferation of hand-held devices such as

computer laptops, tablets, and smart phones, there is

increased easy access to online resources and video

games. There is a body of research that points to

unique learning habits of young people who pre-

fer short visual explanations, to receive information

quickly, prefer multi-tasking and non-linear access to

information, have a low tolerance for lectures, pre-

fer active rather than passive learning, and are ki-

naesthetic, experiential, hands-on learners who must

be engaged with ﬁrst-person learning, games, sim-

ulations, and role-playing (Junco and Mastrodicasa,

2007); (Oblinger and Oblinger, 2005); (Tapscott,

2009). Although playing computer and video games

are largely seen as a distraction to learning, they are

recognized as valid cognitive activities since they af-

∗

This research was supported by the Social Sciences

and Humanities Research Council (SSHRC) and FED

DEV/OBI and NSERC grants of Canada.

fect a player’s ability to self regulate, make right de-

cisions, and problem-solve ((Dance, 2003), p. 177).

The goal of this research is to use already avail-

able technology devices (computers and online video

games) accessed by youth to identify children with

learning differences that may be affecting their learn-

ing abilities (e.g., the acquisition, retention, under-

standing, organization of information). This re-

search discusses an extension to an earlier system

(Whent et al., 2012) for identifying a child’s cognitive

skill level called “Think-2-Learn”(presently renamed

“Thriver”), created by OTEP Inc. (Online Training &

Evaluation Portal).

Section 2 presents the related work on computer

gaming, cognition and learning, OTEP’s Thriver data

warehousing and mining approach, and on some

existing data warehousing schema integration ap-

proaches. Section 3 presents the new additional au-

tomatic data warehouse integration approaches being

proposed for advancing the OTEP solution, including

the automatic data warehouse schema integration al-

gorithm, automatic querying and data cleaning algo-

338

I. Ezeife C., Whent R., Martinovic D., Frost R., Alahmad Y. and Mumu T..

Automatic Updating of Computer Games Data Warehouse for Cognition Identiﬁcation.

DOI: 10.5220/0004849203380345

In Proceedings of the 6th International Conference on Computer Supported Education (CSEDU-2014), pages 338-345

ISBN: 978-989-758-020-8

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

rithms. Section 4 discusses performance analysis of

how games data can be effective in identifying cogni-

tion while section 5 presents conclusions and future

work.

2 RELATED WORK

2.1 Computer Games, Cognition and

Learning

Positive effects of playing computer games on cog-

nitive development include improving visual intelli-

gence skills which are useful in science and technol-

ogy, and generally in such ﬁelds requiring manipu-

lation of images on a screen (Subrahmanyam et al.,

2000). Although other consequences of playing com-

puter and video games have been studied (Marti-

novic et al., 2011) computer-based games may en-

hance hand-eye coordination, visual scanning, au-

ditory discrimination, and spatial skills (DeLisi and

Wolford, 2002). It has been stated that repetitive

game playing may increase young children’s working

memory (Thorell et al., 2009), mental rotation accu-

racy (DeLisi and Wolford, 2002), and spatial rotation,

iconic skills, and visual attention (Subrahmanyam

et al., 2001). Playing the carefully and purpose-

fully designed computer games may positively affect

learning among children of wide range of ages (Sub-

rahmanyam et al., 2000), (Martinovic et al., 2014a),

(Martinovic et al., 2014b). This is because playing

computer games involves integration of “touch, voice,

music, video, still images, graphics, and text” (IBM,

1991), and can stimulate a variety of intelligences

(e.g., linguistic, logical, spatial, kinaesthetic, musi-

cal), that may particularly inﬂuence development of

literacy skills and ability to problem-solve.

2.2 Data Warehouse Schema

Integration Approaches

A data warehouse is a historical, integrated, subject-

oriented database storing data from multiple data

sources in the one data warehouse schema (Han et al.,

2011). Construction of a data warehouse is done

through processes of schema and data integration of

different data sources which involve data cleaning

(Ezeife and Ohanekwu, 2005), data transformation

and loading with periodic refreshing. A popular data

warehouse schema approach is the star schema where

there is a central fact table having foreign key at-

tributes that include the main subjects of interest, the

integration attribute, the historical time attribute and

some non-foreign key aggregate measures of inter-

est. Other descriptive tables in the data warehouse

design using the star schema are dimension attributes

for describing the foreign key attributes in the fact ta-

ble (Ezeife, 2001). A measure such as score achieved

during a game by a child can be calculated from

a multidimensional model version of the data ware-

house called the data cube (Ezeife, 2001). Existing

schema integration approaches (Kern et al., 2011),

(Fan and Poulovassilis, 2004), (Rahm and Bernstein,

2001) process some common steps during schema

integration. The ﬁrst step in integrating schemas

(e.g., integrating schema1(Cust, C#, CName, First-

Name, LastName) and schema2(Customer, CustID,

Company, Contact, Phone)) is to identify and char-

acterize these inter-schema relationships between the

multiple data source schemas to be integrated. This

schema element relationships can be identiﬁed auto-

matically through integration approaches such as ap-

plication domains, match operator, architecture for

generic match, schema-level matchers, instance-level

approach. Application domains approach integrates

an independently developed schema with a given con-

ceptual schema and requires semantic query process-

ing. Once these schema relationships are identi-

ﬁed, matching elements can be uniﬁed under a co-

herent, integrated schema or view. Match operator re-

quires a representation for its input schemas and out-

put mapping and needs to explore many approaches

to match. For example, the result of calling Match

on the two schemas above could be Cust.C# = Cus-

tomer.CustID, Cust.CName = Customer.Company,

and {Cust.FirstName, Cust.LastName} = Cus-

tomer.Contact. Schema-level matchers only consider

schema information, not instance data. The available

information includes the usual properties of schema

elements, such as name, description, data type, re-

lationship types (part-of, is-a, etc.), constraints, and

schema structure. Instance-level data can give impor-

tant insight into the contents and meaning of schema

elements. For example, a data-guide or an approx-

imate schema graph can be generated automatically

from XML documents. In (Kern et al., 2011) a frame-

work for building logical schema for federated data

warehouse from different data warehouse resources

was proposed. The logical schema of the federated

data warehouse is generated as a result of integration

of components of data warehouses (the fact and di-

mension tables). The input to the integration process

consists of several sets of fact table (F) with dimen-

sion tables (Dim) that are related to the fact table F

through foreign key constraints. To integrate the data

warehouses into one federated data warehouse, the al-

gorithm begins with an empty fact table (F

output),

AutomaticUpdatingofComputerGamesDataWarehouseforCognitionIdentification

339

and for each measure aggregate attribute in an input

fact table F input, it looks for a corresponding mea-

sure attribute and if it exists, it deﬁnes the mapping

between these two attributes in the input and output

fact table. If it does not exist in the output fact ta-

ble, then, the new measure attribute is inserted in the

output fact table and the mapping between these input

and output fact table measure attribute is deﬁned and

inserted. Then, for each dimension table in the input

data warehouse, D input, it matches the foreign key

attribute of the table with those of the output dimen-

sion table if this input dimension table exists in the

output data warehouse and deﬁnes the mapping be-

tween this input dimension table and the output fact

table. If the input dimension table does not exist in the

output data warehouse, or none of its attributes match

in schema F output, then it adds the dimension table

to federation schema F output. In (Fan and Poulo-

vassilis, 2004), a heterogeneous data transformation

and integration system, named AutoMed, that offers

the capability to handle data integration across multi-

ple data models and supports a low-level hypergraph-

based data model (HDM) was proposed. For any

modeling language M, data source wrappers translate

data source schemas expressed in M into their Au-

toMed representation, and for every construct of M

there is an adds and a deletes primitive transforma-

tion which add to/delete from a schema an instance of

that construct.

2.3 OTEP Data Warehouse Integration

Approach

Whent et al. (2012) presented the OTEP system

which uses online games to screen or assess children’s

cognitive skills in order to later suggest a learning

plan that would be most suitable for their learning

success. Thus, the paper described an approach for

gathering and integrating the relevant data from (1)

video games data, (2) cognitive skills and mapping

data and to obtain a data warehouse schema called

OTEP GamesDW. The input games data source that

was integrated had then, 100 games that a child can

play. The games data source containing information

about each game, user’s record of game plays, user

information, game categories etc. were represented in

about eight database tables. The second data source

integrated into the OTEP GamesDW is the cognitive

data source, which describes the cognition levels and

their connections to the game instances in the ﬁrst

data source. The system used about 10 main cogni-

tive categories such as Visual Processing, Processing

Speed, Auditory Processing, etc., and two to eight sub

cognitive categories (e.g., verbal output and written

output as sub categories of processing speed). The

integrated data warehouse has the following fact ta-

ble with attributes from games data source and cogni-

tive data source. FactTable(userid, gamid, gameseq,

gameDB, gamelevelid, catid, normcogid, cogid, cog-

subid, time-m, coglevel, gamescore, duration, tries);

This fact table along with accompanying dimension

tables can be used to answer queries like What are

the cognitive norms (based on cognitive categories at-

tached to the games) and game achievement norms

(based on the average game play scores) for chil-

dren who are 8 years old and who have difﬁculty

reading for a reading game?. Currently, the schema

integration is typically performed manually, perhaps

supported by a graphical user interface, that is a te-

dious, time consuming, error-prone, and therefore ex-

pensive process. To provide automated support suit-

able for integrating new changes in the data sources as

well as integrating new data sources such as those for

connecting learning with both cognitiveachievements

and games play achievements, we proposed a generic,

customizable implementation of the Match operator

that is usable across application areas which makes it

easier to build application-speciﬁc tools that include

automatic schema match. Our proposed OTEP auto-

matic integration approach is based on combining the

application domains and schema-level matchers.

3 THE AUTOMATIC OTEP

DWH

SCHEMA GENERATION

APPROACH

3.1 The Automatic OTEP Model and

Problem Addressed

In the ﬁrst phase of OTEP Inc. (Online Training &

Evaluation Portal) project (Whent et al., 2012), the

data warehouse schema was built manually by the

developers. Continuous manual data warehouse in-

tegration is tedious and time consuming because the

database developer or the administrator spends a lot

of time creating the initial schema of the data ware-

house. In addition, there is need to keep monitor-

ing any changes in all the corresponding database

sources, or to integrate new data sources such as new

games sources or learning data source to reﬂect the

change and update the data warehouse schema. Thus,

to have a more correct, effective and available data

warehouse structure, this paper proposes advancing

the initial OTEP system with the ability to do auto-

matic data warehouse integration and refreshing to

accommodate new changes in source schemas, or in-

CSEDU2014-6thInternationalConferenceonComputerSupportedEducation

340

tegrate new schemas. It also proposes an automatic

querying interface for online analytical processing.

The existing OTEP model (Whent et al., 2012)

measures a child’s cognitive abilities through his/her

performances in repetitive playing of a variety of

games in different cognitive categories. The model

accomplishes this goal by comparing the child’s

performance in these games with the performances

of dynamically changing normalized performances

(termed norms) of other children in similar compar-

ison groups such as age, ethnic background, social

background, learning or physical, etc. Thus, OTEP

system uses data warehouse integration approach to

integrate game playing database, cognitive inventory

database, and other data sources such as learning

inventory database and online analytical processing

(olap) approach with multidimensional views (Ezeife,

2001) as well as data mining approaches for querying.

The game playing database can also result from a con-

tinuous integration of various gaming sites.

3.2 The Automatic OTEP Data

Warehouse(OTEP

DW auto)

Algorithms

The goal of this system is to automatically build, re-

fresh and update the integrated, historical data ware-

house of online games play records of children,

their cognitive and learning characteristics. These

data warehouses are used to screen or assess chil-

dren’s cognitive skills and later suggest a learning

plan that would be most suitable for their learn-

ing success. This paper describes the algorithms

for automatically integrating the relevant data from

(1) video games data, (2) cognitive skills and map-

ping data and to obtain a data warehouse schema

called OTEP

GamesDW auto. In the future, other

data sources will be integrated including the learn-

ing achievement data and third party data. The cur-

rent schemas of the games data source and the cog-

nitive data source with the integrated data warehouse

are provided in this section. Three automatic algo-

rithms for schema generation, view (querying) gener-

ation and data cleaning are presented.

3.2.1 The OTEP DW auto Schema Generation

Algorithm

The input of the OTEP

DW auto schema generation

algorithm is the Database Name (e.g., Thrivergames,

Discovery which are names of the database to be in-

tegrated automatically) which contains the connec-

tion parameters. After the connection to the database,

the algorithm queries the table name sourceStructure

Table 1: A Segment of the sourceStructure Metadata of Two

Data Sources.

ﬁeld ﬁeld ﬁeld ﬁeld ﬁeld ﬁeld ﬁeld

Name Type Size Const Source Table Table

raint Name Type

user Num 20 prima Think2 wp t2l dimen

id ber ry key Learn user sion

user Var 20 unique Think2 wp t2l

score Num 10 Think2 wp t2l aggre

ber Learn game log gation

dura Num 10 Think2 wp t2l aggre

tion ber Learn game log gation

which contains metadata information (consisting of

all the attributes (ﬁelds) in the databases and their de-

scriptions) about the two database sources. An ex-

ample schema for the metadata table, sourceStruc-

ture is sourceStructure (ﬁeldname: string, ﬁeldType:

string, ﬁeldSize: integer,ﬁeldConstraint: string, ﬁeld-

Source: string, ﬁeldTableName: string, ﬁeldTable-

Type: string). The description of the steps in the

proposed Automatic OTEP

DWH schema generator

algorithm are presented next. Step 1: If no data ware-

house, called OTEP

DWH already exists in the server,

then create an empty data warehouse structure called

OTEP

DWH. Otherwise go to Step 9.

Step 2: Sort the table sourcesStructure which is given

as input to the algorithm by attribute ﬁeldTableName.

Table 1 shows an example, illustrating the structure

and contents of table sourcesStructure.

Step 3: Read all the attributes of the table (from ﬁeld-

Name of sourceStructure table) in the database for

purposes of mapping to an existing attribute or adding

to the existing schema.

Step 4: In this step the algorithm creates all the di-

mension tables of the data warehouse. The algo-

rithm sequentially reads the value of attribute ﬁeld-

Name from sourcesStructure table as per step 3. It

reads the table name of that attribute from ﬁeldTable-

Nameattribute as in step 4.1. It checks whether the

table name is marked as dimension table in attribute

ﬁeldTableType as per steps 4.2 to 4.5. If the table

name already exists in OTEP

DWH schema, then the

algorithm adds the new attribute and maps it to the

related dimension table. If the table does not exist in

OTEP

DWH schema, then the algorithm creates the

dimension table with the name of the value of ﬁeld-

TableNameattribute concatenated with

dim string (to

distinguish the dimension tables), the new created di-

mension table including the attribute as per step 4.4.

The algorithm iteratively repeats step 4 for each at-

tribute its tables marked as a dimension table until it

builds all the dimension tables.

Step 5:Create the fact table named factTable. The

fact table represents the central table of the star

schema with major subject, integrated, and historical

AutomaticUpdatingofComputerGamesDataWarehouseforCognitionIdentification

341

attributes. For each primary attribute in dimension ta-

bles, the algorithm adds the attribute to the fact table

as a reference (foreign key) attribute which refers to

the dimension table.

Step 6:add the subject attributes which have a value

subject in attribute ﬁeldTableType to fact table fact-

Table.

Step 7:Add the integration attribute to the fact table

factTable. The integration attribute is used to distin-

guish the source of the database from which the orig-

inal record was fetched.

Step 8: Add the historical attribute to the fact table

factTable. The proposed algorithm adds the attribute

name dateTime which stores the date and time of cre-

ation of the record in the fact table, factTable.

Step 9: Extract all table names and attributes from the

source structure table that have ﬁeldTableType value

as dimension.

Step 10: For each dimension table, match the table

name with given remote database tables to be inte-

grated in the existing data warehouse DWH. We de-

ﬁne the match operator with the keyword

{%users%, %games%,%collections%,%level%},

and each keyword has subk-keywords

for example user keyword has subkeywords

{%info%,%profile%}.

Step 11: For each matched table, extract the

data of all the attributes having primary key, and the

data into the fact and corresponding dimension tables.

3.2.2 The OTEP DW auto View Generation

Algorithm

In this phase of the OTEP project, a dynamic graphi-

cal user interface (GUI) which allows the end user to

query and browse the contents of the data warehouse

in different views was built. The interface is user

friendly and has the ﬂexibility to compose any kind

of query on the data warehouse presenting the result

as a view. The following are some queries that can

be answered as views by the data warehouse. For

this reason, we propose an algorithm to automatically

generate the required view by the end user. Algorithm

1 shows the automatic view generator algorithm. Q1)

list all students with their ages, source database, and

number of played games for all the periods.

Q2) for a given student ID, list all the played games

by the student including the completed levels,

achieved score, number of tries, and duration.

Q3) for a given student ID, list all played games by

student including respective main-category, respec-

tive sub-category, score, and derived performance.

Q4) view the matrix performance for a given student

in each individual model.

Q5) view the required performance, achieved perfor-

mance in a speciﬁc cognitive skill with a speciﬁc/all

cognitive main category in a speciﬁc model for a

given student.

Q6) For a given student ID, list all the played games

by the student including score, compared to highest

score, lowest score, and norm among all students

who played the same games.

Algorithm 1. (The Automatic OTEP View Generation Al-

gorithm.)

Algorithm OTEP

DWHview ()

Input: list of all parameter attributes ﬁelds

(columns need to be shown in query result ),

condition criteria

Output: view containing the execution result of the

query

BEGIN

1. Find the table name in OTEP

DWH data warehouse for

each attribute in ﬁelds input

2. Create data query language (DQL) as a select statement

3. Concatenate the input condition to the query statement

4. Submit the query to the OTEP

DWH data warehouse

and store the result

5. Return the result to the end user

END

3.2.3 The OTEP

DW auto Data Cleaning

Algorithm

Our extraction system faced a lot of challenges dur-

ing extraction of data contents from each individual

video game database source in order to load them into

OTEP

DWH. This is because of the different formats

and representation of the games data. In addition, it

is due to the existence of different database schemas

and structures. Thus, we implemented a cleaner

algorithm shown as 2 for cleaning and extracting the

data of interest and loading it in the right position

in data warehouse. The OTEP system also keeps

track of all the new modiﬁcations such as add, delete,

modify table or attribute in database source.

Algorithm 2. (The Automatic OTEP Data Cleaning Al-

gorithm.)

Algorithm OTEP

DWHcleaner ()

Input: data records of all database sources

Output: clean data loaded in data warehouse

BEGIN

1. Remove the white spaces of those data records have

value of attribute gameStatus equals to completed

2. Remove all special characters and symbols such as

{,}, /,[,],<,>,,;

3. For each attribute name located in OTEP

DWH, extract

the next token which represents the value of attribute in

CSEDU2014-6thInternationalConferenceonComputerSupportedEducation

342

the right position in data warehouse

4. For the user

and age of the student because the value of the attribute

is given in the format such as 111111

M 14 this means

that the studentID=111111 is male (M) and 14 years old

age.

END

4 PERFORMANCE STUDY AND

USE OF OTEP SYSTEM

The goal of this paper is to propose computer sci-

ence automatic data integration methods (algorithms)

that are used to extend the OTEP system for identi-

fying cognition through repetitive video game play-

ing. Thus, a performance comparison is one which

shows that the extensions provided by the new sys-

tem are correct and more effective in integrating more

data sources automatically and handling more com-

plex queries. We shall provide in the next subsections

discussions of the extensions provided by the system

and details of how to use the OTEP system to cor-

rectly identify cognition.

4.1 Correctness of Extensions Provided

by New OTEP System

An example automatic integration performed with

our extended OTEP system has the ability to inte-

grate more than one cognitive skills matrix model

for map the video games performances of a player

to cognitive skills levels. In the earlier OTEP sys-

tem (Whent et al., 2012), only one cognitive matrix

model, the Crouse model (Crouse, 2010) was used

while the current system proposed in this paper al-

lows integration of more than one model now includ-

ing also using the Reed cognitive matrix model (Mar-

tinovic et al., 2014a). Each of the cognitive mod-

els provides both the cognitive classiﬁcation model

(called the cognitive matrix) and the cognitive corre-

lation matrix (called skills matrix). The cognitive ma-

trix model provides a method for classifying simple

responsible video games into one of the main cogni-

tive categories and subcategories. The cognitive skills

matrix speciﬁes the correlation between areas of cog-

nitive processing and student achievement. For ex-

ample, with the Crouse model, there are 6 main cog-

nitive categories (such as auditory, visual, sequential

rational, concept, speed and executive) and 2 to 8 sub

categories such as (short-term memory for visual de-

tails, talking speed, etc.). The games in our repository

are classiﬁed into a main cognitive category and sub-

categories so that our integrated data warehouse sys-

tem can be used to gather for each player, the histor-

ical game play data such as scores achieved in each

game, number of trials for each game level and the

time needed to complete each game level. Our sys-

tem computes the game play norm (average as norm

and/or any other measures such as variance, standard

deviation) of a comparison group (e.g., all 8 year old,

all male players, etc.) so that the performance of the

player is compared with this norm and their cognitive

level could be identiﬁed with the cognitive correla-

tion matrix model (called skills matrix) using a model

such as Crouse’s or Reed’s. The cognitive skills ma-

trix speciﬁes the correlation between areas of cog-

nitive processing and student achievement. For ex-

ample, with the Crouse model used in (Whent et al.,

2012), it is indicated that for cognitive skill of basic

reading, in the 6 main cognitive categories of audi-

tory, visual, sequential, concept, speed and executive,

a player’s basic reading skills is taken to be good if

their computed game play record in auditory games

is high (as determined using the bell shape and the

norms and the standard deviations), visual is moder-

ate, sequential is high and speed is high. The newly

integrated Reed’s model consists of 9 main cognitive

categories and 43 cognitive subcategories.

Another example is that the existing system had

been extended with this approach to move from

100 video to about 200 video games in its reposi-

tory. Other usability features added include automatic

querying capabilities with automatic views for a wide

range of cognition-related queries.

4.2 How to Use the OTEP System

In our research we work with simple, single-player

games that potentially target and measure the key cog-

nitive skills in children and adolescents. In addition

to carefully analyzing each game, we also look into

the player’s performance (e.g., time spent on task,

repetition of trials, engagement, and use of hints),

note the background information (e.g., grade level,

age and gender), and acquire input from their par-

ents and teachers. The list of cognitive skills is

presented in a cognitive matrix at the Online Train-

ing & Evaluation Portal (OTEP Inc., Whent et al.,

2012); it has 9 main cognitive categories (visual per-

ception, visual attention, visual motor, auditory pro-

cessing, executive function, memory, acquired cogni-

tion, social cognition, and emotional cognition) and

43 sub-categories (e.g., visual tracking, selective at-

tention, problem solving, and semantic memory). Our

interdisciplinary team uses two web sites; the por-

tal for parents, Thriver.ca, and a site for gamers,

ThriverGames.ca, which currently has 167 games.

AutomaticUpdatingofComputerGamesDataWarehouseforCognitionIdentification

343

The games are grouped according to the cognitive

skills they employ, based on the cognitive matrix.

This classiﬁcation helps us to determine in which

categories we are still missing games, and can also

be used when suggesting to children which games

to play next. We also invite a child’s parents or

caregiver to complete a survey about his/her learn-

ing style, behaviour, and his/her cognitive strengths

and weaknesses. The survey and gaming informa-

tion are recorded in a database suitable for search-

ing and retrieving data, and producing reports. An

enhancement of the software system (the web sites

and a database) will use these data to create a per-

sonalized plan for the child with recommendations of

which games to play next and other strategies that the

whole family can use to support their child’s cogni-

tive development. These recommendations are based

on our extensive literature reviews that are ongoing

and will continue throughout this project. This sys-

tem could be used under a variety of conditions (e.g.,

in school or at home), could be designed to provide

feedback to the child, parent, or professional (e.g.,

teacher,psychologist), and could work under different

models (e.g., behaviourist or cognitive model), based

on the parameters selected by the user.

Our target population are 7-12 year olds and their

parents/caregivers. There are three ways in which one

could participate in our study: (a) as an online games

player (contributing to a pool of normative gaming

data, based on the playerage, grade level, and gen-

der); (b) as a parent, who completes an online survey

and enrolls the child to play games online (where the

survey and gaming data are triangulated and a child’s

cognitive proﬁle is created); (c) as a face-to-face par-

ticipant in our controlled lab environment (where par-

ents complete the survey and children complete cog-

nitive and academic achievement tests, and are ob-

served during selected game play to record engage-

ment in gaming). Presently we are still collecting data

in a controlled environment in which the child does

NEPSY II (Pearson) test and plays 15 games that tar-

get the comparable cognitive skills as NEPSY II.

So far we have extended OTEP’s repository of

simple online computer games to 167 games and vali-

dated these games to determine (according to the cog-

nitive matrix), the primary and secondary cognitive

skills engaged in the players during each game (Mar-

tinovic et al., 2014a). We worked in parallel on: a)

establishing a literature review in the area of play-

ing simple computer games (i.e., single-player games

that are relatively short and are high in activation of

speciﬁc cognitive processing domains) and their rela-

tion to cognitive effects/gains among children, while

putting speciﬁc emphasis on a design, reliability and

validity of instruments and methods used; and b) in-

vestigating the feasibility of various methods for eval-

uating the relationship between games and children’s

cognitive skills. Based on our present data collection

in the lab environment, we intend to establish correla-

tion between: on one side–the child’s cognitive skills

and learning style, and on the other side–the child’s

games play data.

5 CONCLUSIONS

In this paper, we presented the extension made to our

current work on the online product called “Thriver”

developed by OTEP Inc. (Online Training & Eval-

uation Portal). OTEP video games source databases

continues to grow and has grown from 100 to about

200 games whose records need to be integrated into

the data warehouses for correct querying. Thus, the

need to build an automatic schema and data integra-

tor, view generator and data cleaner for continuous in-

tegration of new games and other data sources into the

system. The OTEP system is intended to record play-

ers’ scores to continuously assess and monitor their

cognitive strengths and weaknesses with regards to

the main cognitive categories. The Web based tool for

identifying cognitive skill level is developed as an in-

tegration or data warehouse of a number of relevant

data sources such as the cognitive skills categories

data (which also changes as provided by the psy-

chologists), games data (changing as more games are

placed in the repository), player inventory data and so

on. The integrated data are continuously mined, ana-

lyzed and queried for proper and quick assessment or

recommendations.

We continue to work on extending the system: (a)

increasing number of games; (b) increasing a relia-

bility of categorization of the games by achieving an

agreement between 2-3 scorers; (c ) new cognitive

matrix categorizations (for example, Crouse’s or any

other.) (d) developing a formula that will incorpo-

rate the features of the game (including differentiat-

ing the impact of different cognitive sub categories),

the number of trials, the scores achieved and the time

spent playing. Future work also include tracking chil-

dren’s cognitive development, proposing remediation

in terms of games that may strengthen some cognitive

abilities, and increasing validity and reliability of our

approach.

CSEDU2014-6thInternationalConferenceonComputerSupportedEducation

344

REFERENCES

Crouse, S. (2010). The Cognitive Processing Inventory

(CPI). LDinfo Publishing.

Dance, F. (2003). The digital divide. In Strate, L., Jacobson,

R. J., and Gibson, S., editors, Communication and cy-

berspace: Social interaction in an electro nic environ-

ment. Hampton Press, Inc., Cresskill, New Jersey.

DeLisi, R. and Wolford, J. (2002). Improving children’s

mental rotation accuracy with computer game playing.

Journal of Genetic Psychology, 163(3):272–282.

Ezeife, C. (2001). Selecting and Materializing Horizon-

tally Partitioned Warehouse Views. Journal of Data

& Knowledge Engineering, 4(1). Elsevier Sciences.

Ezeife, C. and Ohanekwu, T. (2005). Use of smart tokens

in cleaning integrated warehouse data. International

Journal of Data Warehousing and Mining (IJDWM),

1(2):1–22.

Fan, H. and Poulovassilis, A. (2004). Schema evo-

lution in data warehousing environmentsa schema

transformation-based approach. In proceedings of

ER 2004 Conference, In Conceptual Modeling, Berlin

Heidelberg, pp 639-653. LNCS, Springer.

Han, J., Kamber, M., and Pei, J. (August 2011). Data Min-

ing: Concepts and Techniques . Morgan Kaufmann

Publishers.

IBM (1991). Literacy development. (ibm’s writing to

read, and principles of adult literacy system multime-

dia/educational software). T H E Journal (Technologic

al Horizons in Education).

Junco, R. and Mastrodicasa, J. (2007). Connecting to the

net.generation: What higher education professionals

need to know about today’s students. Student Affairs

Administrators in Higher Education (NASPA), Wash-

ington, DC.

Kern, R., Ryk, K., and Nguyen, N. T. (2011). A frame-

work for building logical schema and query decom-

position in data warehouse federations. In In Compu-

tational Collective Intelligence. Technologies and Ap-

plications, pp 612-622. LNCS, Springer.

Martinovic, D., Ezeife, C., Whent, R., Reed, J., Burgess,

G. H., Pomerleau, C. M., and Chaturvedi, R. (2014a).

Critic prooﬁng of the cognitive aspects of simple

games. Computers and Education, 72:132–144. .

Martinovic, D., Freiman, V., and Karadag, Z. (2011). Child

and youth development beyond age 6 - transitions to

digitally literate adults. Unpublished Report. Ministry

of Child and Youth Services.

Martinovic, D., Whent, R., Adeyemi, A., Yang, Y., Ezeife,

C., Lekule, C., Pomerleau, C., and Frost, R. (2014b).

Gamiﬁcation of life: Playing computer games to

learn, train, and improve cognitively. Journal of Edu-

cational and Social Research. Mediterranean Center

of Social and Educational Research, Rome, Italy(to

appear).

Oblinger, D. and Oblinger, J. (2005). Ed-

ucating the net generation. educause.

net.educause.edu/ir/library/pdf/pub7101.pdf.

Rahm, E. and Bernstein, P. A. (2001). A survey of ap-

proaches to automatic schema matching. The VLDB

Journal, 10(4):334–350.

Subrahmanyam, K., Greenﬁeld, P., Kraut, R., and E.Gross

(2001). The impact of computer use on children’s and

adolescents’ development. Journal of Applied Devel-

opmental Psychology, 22(1):7–30.

Subrahmanyam, K., Kraut, R., Greenﬁeld, P., and Gross, E.

(2000). The impact of home computer use and chil-

dren’s activities and development. The Future of Chil-

dren: Children and Computer Technology, 10(2):123–

14.

Tapscott, D. (2009). Grown up digital: How the net gen-

eration is changing your world. McGraw Hill, New

York, NY.

Thorell, L., Lindqvist, S., Nutley, S., Bohlin, G., and Kling-

berg, T. (2009). Training and transfer effects of exec-

utive functions in preschool children. Developmental

Science, 12(1):106–113. .

Whent, R., Martinovic, D., Ezeife, C., Ahmed, S., Ah-

mad, Y., and Mumu, T. (2012). School age chil-

dren’s cognition identiﬁcation by mining integrated

computer games data. In proceedings of the 4th In-

ternational Conference on Computer Supported Edu-

cation (CSEDU 2012), Porto, Portugal, pp 495-505.

SciTePress.

AutomaticUpdatingofComputerGamesDataWarehouseforCognitionIdentification

345