Mouhsine Elmoudir

IDSD

THE PRINCIPLE OF DATA MINING

A. CHAPTER 1 : we see in this explanation the definition of data mining
major , start by the KDD (Knowledge discovery of data) process which
maintain the Data Sources and transform it to Data Store by using Data
integration , so as to move the data preparation and then use directly
the data mining to gain the patterns and finally arrive to the knowledge
needed.
-Moving to the datasets which has examples (instances) , we can see the
labelled data (with target) or unlabelled data (without target)
-In the main of data mining there’s two types of machine learning ,
supervised learning which includes classification, regression, association
rules… .And the unsupervised learning contains the clustering(K-means)

B. CHAPTER 2 : This chapter shows us how data can effect the data
mining : type of variable -> nominal variables, binary variable, ordinal
variable…

-Also , there’s the categorical and continuous attributes so as to

preprare to a very important part called the data preparation and data
cleaning without mess to remove the missing values
- So, it introduces the standard formulation for the data input to data
mining algorithms that will be assumed throughout this book. It goes on
to distinguish between different types of variable and to consider issues
relating to the preparation of data prior to use, particularly the presence
of missing data values and noise. The UCI Repository of datasets is
introduced

C. CHAPTER 3 : we will discover in part an supervised learning

techniques. But first by giving the meaning of classification which looks
like many practical decision-making tasks can be formulated as
classification problems... assigning people or objects to one of a number
of categories, for example customers who are likely to buy or not buy a
particular product in a supermarket.
-We have also the Naïve Bayes algorithm known as probability theory to
find the most likely of the possible classifications. And we can use in
everyday life: the probability of an event, e.g. that the 6.30 p.m. train
from London to your local station arrives on time, is a number from 0 to
1 inclusive, with 0 indicating ‘impossible’ and 1 indicating ‘certain’.
-So, the Naïve Bayes algorithm gives us a way of combining the prior
probability and conditional probabilities in a single formula, which we
can use to calculate the probability of each of the possible classifications
in turn. Having done this we choose the classification with the largest
value
-We discussed also the K-nearst neighbours algorithm which is mainly
used when all attribute values are continuous, although it can be
modified to deal with categorical attributes. The idea is to estimate the
classification of an unseen instance using the classification of the
instance or instances that are closest to it, in some sense that we need
to define and move to the next steps is to find the k training instances
that are closest to the unseen instance. Then take the most commonly
occurring classification for these k instances