Enriching What-If Scenarios with OLAP Usage Preferences

Mariana Carvalho and Orlando Belo

ALGORITMI R&D Centre, Department of Informatics, School of Engineering, University of Minho,

Campus de Gualtar, 4710-057, Braga, Portugal

Keywords: Business Intelligence, What-If Analysis, on-Line Analytical Processing, Usage Preferences,

Multidimensional Databases.

Abstract: Nowadays, enterprise managers involved in decision-making processes struggle with numerous problems

related to market position or business reputation of their companies. Owning the right and high quality set of

information is a crucial factor for developing business activities and consequently gaining competitive

advantages on business arenas. However, retrieving information is not enough. The possibility to simulate

hypothetical scenarios without harming the business using What-If analysis tools and to retrieve highly refined

information is an interesting way of achieving such advantages. In this paper, we propose an approach for

helping to optimize enterprise decision processes using What-If analysis scenarios combined with OLAP

usage preferences. We designed and developed a specific piece of software, which aims to discover the best

recommendations for What-If analysis scenarios’ parameters using OLAP usage preferences, which

incorporates user experience in the definition and analysis of a target decision-making scenario.

1 INTRODUCTION

What-If analysis (Golfarelli et al., 2006) processes

allow for creating simulation models aiming to

explore the behavior of complex business systems.

More pragmatically, they contribute for analyzing the

effects on the behavior of a particular business system

caused by a change of variables and values, which

usually cannot otherwise be discovered by a historical

data manual analysis process (Koutsoukis et al.,

1999). In a real system, the main advantage of

creating a simulation model through What-If analysis

is to be able to implement changes in characteristics

of the business without endangered it (Kellner et al.,

1999). What-If analysis techniques are one of the

most recently ways to achieve these goals. Decision-

makers need to create What-If analysis scenarios to

test and validate their business hypothesis and support

their decisions. In fact, What-If analysis can be the

safer solution towards some doubt and the decision

maker needs to assess to ensure, if possible, that the

subsequent decision will have some success.

Moreover, it allows for analyzing different scenarios

and perspectives of business, anticipating some

solutions.

Online Analytical Processing (OLAP)

(Harinarayan et al., 1996) systems are one of the most

predominant tools for decision-support systems. They

provide specialized means for business analytics as

well as multidimensional view over business data that

are very efficient logical ways for analyzing

businesses activities and organizations. A decision-

support analysis process is an interactive exploration

of multidimensional databases, often performed in ad

hoc manner that allows for users to see data from

different perspectives of analysis. Decision-makers

frequently post complex queries to OLAP systems,

which originate answers containing huge volumes of

data that are quite difficult for analysis and

consequent usage on business scenarios. Thus, it is

essential to filter this information in a way that

contents do not loose significance, being adjusted

according to users needs and business requirements.

The extraction of usage preferences according to each

analytic session promoted by users may come as an

advantage to decision-makers, since it provides a very

effective way to personalize analytical sessions and

multidimensional data structures acting as their

decision-making support. Currently, OLAP systems

technology already provides means for doing

interactive analysis of multidimensional data based

on a set of navigational operations usage and patterns

(Jerbi et al., 2009).

In this paper we present and discuss the

integration of OLAP usage preferences in

conventional What-If scenarios as a mean to improve

Carvalho, M. and Belo, O.

Enriching What-If Scenarios with OLAP Usage Preferences.

DOI: 10.5220/0006040402130220

In Proceedings of the 8th International Joint Conference on Knowledge Discovery, Knowledge Engineering and Knowledge Management (IC3K 2016) - Volume 1: KDIR, pages 213-220

ISBN: 978-989-758-203-5

213

the quality and effectiveness of decision models

related to problems requiring perceptions from the

user point of view. This will avoid the lack of

expertise of the user in the implementation of What-

If scenarios and models. Thus, we designed and

implemented a decision-support system with the

ability to receive a What-If scenario supported by

analytical preferences, which provides us strong

arguments for improving the simulation of a given

system’s behavior based on the preferences of its

users. The system has the ability to suggest OLAP

preferences, providing to the user the most adequate

scenario parameters according to its needs. It was a

way that we find to enrich and make more valuable

What-If scenarios.

This paper is organized as follows. In section 2 we

present an overview about OLAP preferences and

applications. Next in section 3, we describe the entire

process: the database used for test case study, how the

association rules algorithm works, a brief formal

description of OLAP preferences and how we extract

the association rules of the data cube. In section 4 we

describe the lasts steps of our methodology: showing

how the What-If scenarios are created and enhanced

using the extracted OLAP preferences, moreover we

show how the application works, showing all the

steps between the extraction of the rules until the

definition of the What-If scenario. Finally, section 5

concludes the paper and discusses some possible

future research directions.

2 RELATED WORK

As stated before, the main purpose of this work is to

improve What-If scenarios using OLAP preferences;

in this section we show how preferences are used to

improve decision-making process, either in relational

databases, in OLAP environments, or even in daily

tasks, like traffic analysis. The research on databases

preferences goes back to Lacroix and Lavency

(1987), which was one of the first works that

presented and discussed a preference mechanism as

an extension of a query language. Later, in Agrawal

and Wimmers (2000) it was proposed a formal

framework for expressing and combining user

preferences to address the problem of the high

quantity of available on-line information. After this,

in Chomicki (2003) it was presented a logical

framework for formulating preferences and

embedding them into relational query languages,

which does not impose any restrictions on the

preference relations, and allows for arbitrary

operations and predicate signatures in preference

formulas. After this, a different approach to database

preferences queries presented in Hadjali et al. (2008)

discussed the way we can deal with preferences in a

logic manner using symbolic weights. At the same

time new approaches emerged, new applications

areas arose as well. See for example the work

presented in Letchner et al. (2006) in traffic analysis

where authors presented a set of methods for

including driver preferences and time-variant traffic

condition for in route planning.

OLAP preferences reflect the most interesting

data that decision-making agents selected and

analyzed in past OLAP sessions, using a specific set

of data cubes during certain periods of time

(Harinarayan et al., 1996). More recently, preferences

capture the attention of many researchers in the field

of databases, especially in the field of on-line

analytical processing, approaching the extraction of

preferences using data mining techniques over users’

MDX queries logs (Aligon et al., 2011) or the

specification of an algebra for OLAP preferences

(Golfarelli and Rizzi, 2009). Meanwhile, in

Giacometti et al. (2009) it was presented a

recommender system for OLAP users having the

ability to recommend to the user discoveries detected

on former sessions and in Ahmed et al. (2012) it was

introduced a new approach for user profile

construction based on the information contained in

OLAP query logs. Two years later, in Varga et al.

(2014), it was proposed another framework, but this

one to support users assisting them generally in the

automation of their activities in the context of the next

generation of business intelligence systems using

query recommendation support. Next, in Marcel

(2014) it was summarized various contributions for

developing user-centric OLAP, focusing on the use of

former queries to enhance subsequent analyses. They

show how it can be used in various query

personalization processes or query recommendation

approaches, which vary in terms of formulation

effort, pro activeness, prescriptiveness and expressive

power. To finish this section, just refer the work

presented in Kozmina (2015), which provided a

method for generation of report recommendations

that takes into consideration the preferences of the

users, and, more recently, the work presented in

Bimonte and Negre (2016) that showed the usefulness

of OLAP recommender systems on decision-making

activities.

KDIR 2016 - 8th International Conference on Knowledge Discovery and Information Retrieval

214

3 EXTRACTING OLAP

PREFERENCES

3.1 A General Overview

As stated before, we want to discover the best

recommendations for What-If analysis scenarios

based on past usage information. The main difference

between our approach and standard What-If analysis

methods is the introduction of a process of extraction

of usage preferences on a business multidimensional

database and using them before the simulation of the

model, which allows for having the basis to predict

the behaviour of a given scenario. The process of

extracting OLAP preferences considers five distinct

phases (Figure 1):

Figure 1: Enriching a What-If scenario.

Thus, firstly we start with a view selection process

over the data cube we selected to support our work.

We start with a small case study, which considers a

single fact table - “Internet Sales” - and related

dimension tables – “Dim Currency”, “Dim

Customer”, “Dim Product”, “Dim Promotion”, “Dim

Sales Territory”, and “Dim Date”. All these data

objects were extracted from the

“AdventureWorksDW2014” (Microsoft SQL Server

Product Samples: Database, 2015), a small data

warehouse provided by Microsoft as a product sample

for Microsoft SQL Server. In the second phase, we

create and analyze a specific data cube using the

previous data objects as an example for future

complex cases. With the reduction of the size of the

cube, the complexity of the database reduced as well,

and the same happens with the time need to extract

the data we use and process it. Next, we create the

mining structure and define the mining model to

support a mining association process that runs in the

third phase over the cube we created (Han, 1997).To

do that we selected the Microsoft Association Rules

algorithm (Agrawal and Srikant, 1994) that comes

with Microsoft Analysis Services. This Apriori based

algorithm fits well on mining processes that involves

recommendation engines or processes for finding

correlations between different attributes in a given

dataset – in our case we have a recommendation

engine for the suggestion of the items that are most

likely to appear together in a particular search of a

What-If scenario. As other Apriori based association

algorithms, we can define the minimum and the

maximum support values to control the number of

item sets that are generated, and we can also restrict

the number of rules that a model produces by setting

a value for a minimum probability. Even in this stage,

all the rules and item sets extracted are stored in the

mining model. An association mining model has a

simple structure organized in two blocks: 1) the

information about itself and its metadata, and 2) a flat

list containing item sets and rules. Item set nodes

include the definition of the item set, the number of

cases that it contains, and other diverse information

for support. In turn, a rule node describes a general

pattern for the association of items. Every node has

detailed information about the item set or the rule that

will be relevant in the next steps of the process. All

this is use for defining the OLAP preferences on the

fourth stage.

Figure 2: Extracting OLAP preferences.

The process of extracting OLAP preferences (Figure

2) starts with the choice of a user preference item

from a list of frequent item sets of the mining model

(ordered by probability). Next, it filters the list of

association rules taking in count the chosen item set,

returning only the association rules having the

support and confidence previously defined and

containing the chosen item set. The returned

association rule list is used then to form the set of

OLAP preferences for the user. This means that the

item sets of the association rules returned as outcome

are suggested to the user as its business preferences

and then they will be taken into consideration in the

What-If scenario as configuration parameters.

Knowing beforehand cube preferences can have a

significant impact on the outcome results of the

Enriching What-If Scenarios with OLAP Usage Preferences

215

analytical system. Using OLAP preferences, it is

possible to provide exactly the most relevant and

useful information to each specific user in a specific

analysis scenario. Aftermath, there is a significant

reduction of the cube implementation costs,

processing time and memory usage. The cube will

include in its structure only the data that match user

preferences and so it will returns only the data that

interest to user. Moreover, the entire analysis process

can be improved. As already noticed, a cube is a very

complex data structure and it can be difficult for an

analyst to acquire the information he want. With a

simple interface having the ability to recommend the

right queries based on the history of past analytical

sessions makes much simpler the process of

extracting information.

3.2 Dealing with Preferences

Now it is time to define formally what is a preference,

introducing it with a simple working example based

on the works presented in (Kießling, 2002; Ore and

Ore, 1962). Thus, given a set of attributes A, a

preference P is a strict partial order defined as P (A,

<P), where <P is an irreflexive, transitive and

asymmetric binary relation <P ⊆ dom(A) × dom(A).

If X <P Y, then ‘Y is preferred to X’. A preference P

= (A, <P) is an irreflexive, transitive and asymmetric

binary relation <P on the domain of values of

attributes set A. Let see how this works. If we

consider to analyze how sales vary with the number

of costumers having children living at home, to set its

preferences a user need to choose one of the elements

included in the set of the frequent item sets. This will

allows for choosing the rules that will be used to set

user preferences. Thus, assuming that the user

chooses “Number Children At Home”, using the

previous defined semantics, we have:

In other words, the attribute “Number Children At

Home” is preferred to the attribute “Marital Status”,

“Gender”, “Yearly Income”, “Number Cars Owned”,

“Birth Date”, and so on. Thus, “Marital Status” is

equivalent to “Gender”; “Gender” is equivalent to

“Yearly Income”, “Yearly Income” is equivalent to

“Number Cars Owned”, and so on. Based on this set

of previous preferences, it is possible to select a set of

association rules that contains the attribute “Number

Children At Home” (Figure 3).

Figure 3: The association rules for a given preference -

‘Number Children At Home’.

Accordingly its own business preferences, the user

may choose N association rules, for example the top

3 association rules (Figure 4) of the previous set

(Figure 3), which will be used later to define his

OLAP preferences. This means that the item sets

contained in the filtered association rules will be

suggested to the user as preferences. For example, if

the returned list of association rules is the list in

Figure 4, the recommendations to the user will are

“Number Children At Home”, obviously, “Birth

Date”, “Yearly Income”, “English Education” and

“Total Children”. After this step, the user chooses the

item sets of his preference that will be used as

configuration parameters in the What-If scenario.

Figure 4: A list of some filtered association rules.

4 ENRICHING What-If

SCENARIOS

Basically, What-If analysis can be described as a data

simulation technique whose goal is to inspect the

behavior of a complex system under some given

hypotheses, usually called as scenario. More

pragmatically, What-If analysis measures how

changes in a set of independent variables impact on a

set of dependent variables with reference to a given

simulation model. The integration of OLAP usage

preferences in What-If scenarios for business analysis

enhancement is not a very common thing.

The main focus of a What-If application is a

simulation model (Figure 5). Commonly, this model

is a representation of a real business model that

usually is organized into several application

scenarios. Each one of these scenarios considers a set

of business variables (the source variables) and a set

of setting parameters (scenario parameters). It is the

user that has the responsibility to edit such variables

and obtain some kind of prediction (a new scenario)

for the previous business application.

KDIR 2016 - 8th International Conference on Knowledge Discovery and Information Retrieval

216

Figure 5: A general overview of a What-If analysis process.

The use of What-If analysis allows for the user to

inspect the behavior of a given complex system. The

implementation of a What-If process provides for

several advantages to a business user. It makes

possible to study the behavior of a system without

building it or creating the circumstances to make it

happen in a real world system, clearly saving time and

reducing costs. Besides that, it provides the means to

modify business variables as needed in order to find

potential unexpected behaviors of the business

system, which gives to business managers the

possibility to be aware of the conditions that may lead

to an erratic system behavior and so create the basis

to avoid it in the future.

In order to improve this process, we use

preferences as recommendations to help enhancing

What-If application scenarios. The use of preferences

in this case can be beneficial. The main advantage is

to become possible to simulate a system behavior

based on past data extracted from OLAP sessions.

Preferences have the ability to recommend to the user

the axes of analysis that are strongly related to each

other, helping him to introduce valuable information

in the application scenario he is building. Preferences

are pieces of data that users give more attention in

OLAP sessions, conditioning their ad hoc analysis

and decision-making tasks. They can be defined

based on historical data provided by a simple business

application or from a more sophisticated piece of

software like a data mining system. Often a

preference reflects also hidden patterns that were

detected in the data set. Using association rules based

on preferences has the advantage that the user does

not need to know the business domain. Preferences

can also helping control over the returned

information, providing access to relevant information

and eliminating the irrelevant one. One may not know

the proportions of the outcome: it may be an empty

result, or an information flooding. Due to this, query

runtime can be enhanced against cases without

preferences. Consequently, in our process, we get

more focused and refined results, which helps both a

user who is not familiar with the business analysis and

an analyst who is familiar with the business modeling

data.

Our process proceeds as is shown in Figure 6.

Firstly, we use an OLAP data cube as input. The input

data will be used to define the application scenario

based on historical data extract from previous OLAP

sessions. Then, we define the scenario settings,

delineating the axis of analysis, the set of values for

analyzing, and the set of values to change according

to the goals defined previously. This last step usually

differs among distinct analysis tools. Then, the What-

If process proceeds with choosing a tool. To run a

simulation model (a scenario based on historical data)

it is required an appropriate tool (a What-If scenario

analysis tool), in order to get a prediction scenario.

The What-If analysis tool calculates and lets the user

to explore and analyze the impact of the changes in

the setting values of the entire application scenario. It

is the user that is responsible to accept the new data

cube, or return to change the settings of the

application scenario and make the changes required

over to the target data.

Figure 6: A general overview of a What-If analysis process.

Figure 7: The user interface of the application for What-If

analysis enrichment.

Now we will present an example of the business

application. As stated previously, the extraction of

association rules and OLAP preferences are used to

define the configuration parameters of the What-If

scenario. As a demonstrative example, we will

consider the following business application scenario:

our main goal is to analyze the evolution of the sales

amount (represented by the Y axis) according to the

number of customers’ children at home and year

(represented by the X axis) – value that is given by

the attribute “Number Children At Home”.

Without OLAP preferences (see points 4 and 5 of

Figure 7), an analyst may select for his application

scenario the attributes “Number Children At Home”,

“Sales Amount” and probably “Calendar Year”. If so,

we got a chart like the one presented in Figure 8. So,

with the results got it in the previous section we

Enriching What-If Scenarios with OLAP Usage Preferences

217

Figure 8: A scenario without OLAP Preferences.

realized that there are several attributes strongly

related to our target attribute – “Number Children At

Home” -, such as ‘Yearly Income’ and ‘English

Education, which means that these attributes would

significantly improve our analysis when toke as

configuration parameters in our business application

scenario.

To support and perform What-If analysis

processes we choose Microsoft Excel, since it allows

for creating PivotTable reports based on OLAP

source data. OLAP PivotTable Extensions is an Excel

add-in, which extends the functionality of

PivotTables on Microsoft Analysis Services

multidimensional structures. Excel can be used as an

OLAP analytical tool to easily analyze and modify

data stored on data cubes. It is possible to modify data

using a PivotTable and to recalculate all data as

necessary, and, if the outcome is acceptable, to

publish all changes so that they are copied into the

OLAP cube. It is this property of Excel that allows for

to do What-If Analysis and to create new application

scenarios with the data that was recalculated.

After choosing the parameters for the What-If

scenario, the user can make some changes – e.g.

increasing the total sales values by 10%. Then,

Microsoft Excel calculates how the new value will

modify the old values, based on the properties of

‘What-If Analysis Settings’. Microsoft Excel allows

for the user to calculate data with changes that were

made manually (the user decides when the changes

are made) or automatically (when each value is

changed), to choose the allocation method (‘Equal

Allocation’ or ‘Weighted Allocation’) and finally to

select the value to allocate - the value entered is

divided by the number of allocations or it is

incremented based on the old value. Next, the new

What-If scenario, the scenario with new calculated

values, will be displayed to the user.

In our study, the application we developed allows

for the user: to create What-If scenarios choosing the

available attributes of his choice (first tab in Figure

7); to consult the mining models’ item sets and

association rules (second tab in Figure 7); and both

options together, which we call the hybrid model

(third tab in Figure 7), which aims to create What-If

scenarios using preferences obtained with the mining

models’ association rules. For example, an analyst

wants to analyze how the sales evolve within a

specific customer profile (target audience), in order to

know how sales may vary with the number of cars

owned by the customer. The extraction of association

rules and the sales preferences of the customers may

show that the number of children at home is strongly

related with customers’ yearly income and education.

And with these three analysis axis (perspectives of

analysis), the What-If scenario would be more

accurate and specific, leading to better results, for

example, when a company manager wants to assign a

specific promotion or discount for a specific target

audience.

Figure 9: Example of a business application scenario in a

MS Excel PivotTable.

The Hybrid tab (Figure 7) is the main point of interest

of our application. It provides users with information

about association rules that were extracted from the

cube structure, creates preferences, and recommends

them to the user, in order to create a What-If

application scenario. In the first step of this process,

the user sets values to filter both the support and

probability of both item set and rules (as was seen

before in step 1). This way, it is possible to refine

users’ preferences, leading to a specific and filtered

outcome. The association rules extracted from the

mining model can also be filtered as the item sets (as

seen in step 3 of the process) and displayed ordered

by decreasing values of probability.

In a later phase, the application suggests a set of item

sets (contained in the chosen rules). The user chooses

the item sets, which will be part of the What-If

application scenario. For example, if the chosen item

sets are “Calendar Year”, “Number of Children”,

“Gender” and “Marital Status”, respectively. the

application creates the (partial) PivotTable presented

in Figure 9. We use this set of attributes, instead of

“Yearly Income” and “English Education” in order to

be easy to understand the What-If charts. The

KDIR 2016 - 8th International Conference on Knowledge Discovery and Information Retrieval

218

Figure 10: A possible business application scenario.

Figure 11: The enhanced What-If scenario.

business application scenario we created allows for

the user to see the current sales amounts and its

correspondent growth over the years (Figure 10).

Consequently, after analyzing the current sales

scenario (Figure 10), the user can assign an intended

new final value and analyze how could have been the

evolution of sales amount, in order to achieve the

desired sales amount values. In this case, the

pretended value is the double the final sales amount.

The What-If scenario that was created (Figure 11)

contains the same axis of the previous scenario, but

the values involved with are slightly different.

Nevertheless, there are some limitations associated to

the application we developed. We used one Excel

function, more specifically the What-If analysis

functions over a PivotTable. This function is

performed using the ‘AllocationMethod’ property

setting to ‘Weighted Allocation’ (instead of ‘Equal

Allocation’ setting). With ‘Weighted Allocation’,

Excel calculates the new What-If scenarios (overall

values of the PivotTable) by increasing or decreasing

the overall values in proportion, for example, if the

new final value is 10% higher than the monthly values

are increased by 10%. This must be improved in the

next version of the application for having a more

effective What-If scenario generation.

5 CONCLUSIONS

In this paper we show how OLAP preferences can

contribute for enhancing a What-If scenario,

improving the quality and effectiveness of decision

models where perceptions from the user point of view

can make the difference in a decision-making

process. We implemented a decision-support system

with the ability to receive a What-If scenario that

incorporates usage analytical preferences for

improving the simulation of a given business

application scenario. The system has the ability

provide to its users the most adequate scenario

parameters according to its needs taking into

consideration a set of OLAP preferences that were

extracted from past OLAP sessions. This contributes

significantly to enrich a make more valuable a What-

If scenario for a particular business domain. At this

point, we believe that the process we followed during

the design and implementation of our system can help

in the evaluation of business scenarios that integrate

process solutions for analytical data exploration

environments. Nevertheless there we also recognized

some limitations that need to be overcome, in order to

make the system more efficient, especially at the level

of the usage of Microsoft Office Excel functions and

within the What-If process itself. Additionally, we

need to free the system from some limitations

imposed by user’s choices done in the most parts of

the What-If process. This is must be avoid, because a

user that has limited knowledge about the business

domain or even about the simulation process to be

implemented influences the entire process negatively,

leading consequently to poor results.

Despite the several advantages of using

preferences, there are some drawbacks related to this

process. Due to the use of an association rules

algorithm, sometimes it is difficult to interpret the

results and so some information may be lost. A user,

who is not aware of this process, can have significant

difficulties for exploring meaningful associations.

Additionally, we can face some difficulties in the

What-If process. In a first stage of the What-If

process, if the goal analysis is not done correctly,

What-If questions and scenarios will be not correctly

defined or the preferences outcome will be not

reliable. Thereafter, performed What-If queries will

be not the most suitable process and thus the obtained

prediction will be different of what is expected as a

normal behavior of a real business system. One way

of avoiding this is to study potential and alternative

application scenarios, in order to take the best

advantages of the What-If scenario analysis tool.

Finally, the What-If Analysis results depend strongly

Enriching What-If Scenarios with OLAP Usage Preferences

219

from the data we want to analyze. If it contain some

errors, which is a very common situation, the result

will not be very useful. In order to overcome this kind

of drawbacks, we mainly aim at restructuring

automatically the What-If scenarios, discarding the

user’s dependency and finding a way of overcoming

the limitation we found in some Excel functions.

REFERENCES

Agrawal, R., Wimmers E., 2000. A framework for

expressing and combining preferences. ACM SIGMOD

Record. 29 (2).

Agrawal, R., Srikant, R., 1994. Fast algorithms for mining

association rules. In Proc. 20th int. conf. very large data

bases, VLDB. 1215, pp. 487-499.

Ahmed, E. B., Nabli, A., Gargouri, F., 2012. Building

MultiView analyst profile from multidimensional query

logs: from consensual to conflicting preferences.’ arXiv

preprint arXiv:1203.3589.

Aligon, J., Golfarelli, M., Marcel, P., Rizzi, S., Turricchia,

E., 2011. Mining preferences from OLAP query logs

for proactive personalization. Advances in Databases

and Information Systems. Springer Berlin Heidelberg.

Bimonte, S., Negre, E., 2016. Evaluation of user

satisfaction with OLAP recommender systems: an

application to RecoOLAP on an agricultural energetic

consumption datawarehouse. In International Journal

of Business Information Systems, 21(1), pp. 117-136.

Chomicki, J., 2003. Preference formulas in relational

queries. ACM Transactions on Database Systems

(TODS). 28(4), pp. 427-466.

Giacometti, A., Marcel, P., Negre, E., Soulet, A., 2009.

Query recommendations for OLAP discovery driven

analysis. In Proceedings of the ACM twelfth

international workshop on Data warehousing and

OLAP.

Golfarelli, M., Rizzi, S. Proli, A., 2006. Designing What-If

Analysis: Towards a Methodology. In DOLAP’06,

Arlington, Virginia, USA, pp. 51-58.

Golfarelli, M., Rizzi S., 2009. Expressing OLAP

preferences. Scientific and Statistical Database

Management. Springer Berlin Heidelberg.

Hadjali, A., Kaci S., Prade H., 2008. Database preferences

queries - A possibilistic logic approach with symbolic

priorities. Foundations of Information and Knowledge

Systems. Springer Berlin Heidelberg. pp. 291-310.

Han, J., 1997. OLAP mining: An integration of OLAP with

data mining.’ In Proceedings of the 7th IFIP.

p. 1-9.

Harinarayan, V., Rajaraman, A. Ullman, J., 1996.

Implementing data cubes efficiently. ACM SIGMOD

Record. 25(2).

Jerbi, H., Ravat, F., Teste, O., Zurfluh, G., 2009.

Preference-based recommendations for OLAP analysis.

Springer Berlin Heidelberg.

Kellner, Marc I., Madachy, R., Raffo, D., 1999. Software

process simulation modeling: why? what? how?.

Journal of Systems and Software. 46(2), pp. 91-105.

Kießling, W., 2002. Foundations of preferences in database

systems. In Proceedings of the 28th international

conference on Very Large Data Bases (pp. 311-322).

VLDB Endowment.

Kozmina, N., 2015. Producing Report Recommendations

from Explicitly Stated User Preferences. Baltic Journal

of Modern Computing, 3(2), pp. 110.

Koutsoukis, N., Mitra, G., Lucas, C., 1999. Adapting on-

line analytical processing for decision modelling: The

interaction of information and decision technologies.

Decision Support Systems. 26(1), pp. 1–30.

Lacroix, M., Lavency P., 1987. Preferences; Putting More

Knowledge into Queries. In VLDB. 87, 1987.

Letchner, J., Krumm J., Horvitz E., 2006. Trip router with

individualized preferences (trip): Incorporating

personalization into route planning. In Proceedings of

the National Conference on Artificial Intelligence.

21(2), pp. 1795.

Marcel, P., 2014. Log-driven User-centric OLAP. In

Information and Communication Technology,

Electronics and Microelectronics (MIPRO), 37th

International Convention. pp. 1446-1451.

Microsoft SQL Server Product Samples: Database (2015)

Available at:

[http://msftdbprodsamples.codeplex.com/]

[Accessed 13 fev. 2016].

Ore, O., Ore, Y., 1962. Theory of graphs (Vol. 38).

Providence, RI: American Mathematical Society.

Varga, J., Romero, O., Pedersen, T. B., Thomsen, C. 2014.

Towards next generation BI systems: The analytical

metadata challenge. In Data Warehousing and

Knowledge Discovery. pp. 89-101.

KDIR 2016 - 8th International Conference on Knowledge Discovery and Information Retrieval

220