municipalities expenditures between 1979 and 1987 represent 256 time
series. If you consider three years such as, for example, 1981,1982 and 1983,
you have 256 simple polygonal chains made of two lines segments. Every
couple of segments can approximate a straight line or a convex downward
(or convex upward) simple polygonal chain. The idea is to fi nd outliers
among the couples of segments that performs in a too much different way
from the other couples. In the washer procedure every couple of segments
is represented by an index and a non-parametric test (Sprent test) is applied
to the unknown distribution of those indices. For implementing washer
methodology you can download an open source R (programming language)
function with a simple numeric example.
3.3 Multiple Model Integration
3.3.1 Data Federation
Data federation is a brand new idea for integration of data from many
diffract sources. Many organizations and companies store their data in
different ways, like transactional databases, data warehouses, business
intelligence systems, legacy systems and so on. The problem arises, when
someone needs to access data from some of these sources [8, 4, 3]. There
is no easy way to retrieve the data, because every storage system has its
own way of accessing it. In order to help getting to the data from many
sources, there are some ways to integrate the data, and the most advanced of
them is data federation. To integrate the data it has to be copied and moved,
because the integrated data need to be kept together. Of course it has its
defects, like the time needed to copy and move the data, and some copyright
infringements during copying. The data also occupied more disk space
than it actually needed, because it was kept in few instances. There were
also some problems with data refreshing, because if there was more than
one instance of the data, only the modifi ed instance was up to date, so all
others instances of the data has to be refreshed. Of course it slowed down
the integration system. In response to these problems, the IT specialists
created a new data integration system called data federation. The idea of
data federation is to integrate data from many individual sources and make
access to them as easy as possible. The target has to be reached without
moving or copying the data. In fact, the data sources can be in any location.
It only has to be online. Also, every data source can be made using different
technology, standard and architecture. For the end user it will feel like one
big data storage system. The data federation supports many data storage
standards. From the SQL relational databases like Mysql, PostgreSQL,
InterBase, IBM DB2, Firebird and Oracle through directory services and
object-based databases like LDAP and OpenLDAP, to data warehouses
Data Preparation 53