Machine Learning: Algorithms and Applications Floriano Zini Free - - PDF document
21/05/12 Machine Learning: Algorithms and Applications Floriano Zini Free University of Bozen-Bolzano Faculty of Computer Science Academic Year 2011-2012 Lecture 11: 21 May 2012 Unsupervised Learning (cont) Slides courtesy of Bing
Floriano Zini Free University of Bozen-Bolzano Faculty of Computer Science Academic Year 2011-2012 Lecture 11: 21 May 2012
Slides courtesy of Bing Liu: www.cs.uic.edu/~liub/WebMiningBook.html
n Basic concepts n K-means algorithm n Representation of clusters n Hierarchical clustering n Distance functions n Data standardization n Handling mixed attributes n Which clustering algorithm to use? n Cluster evaluation n Summary
n The distance functions we have seen are for
n In many practical cases data has different types
q interval-scaled q ratio-scaled q symmetric binary q asymmetric binary q nominal q ordinal
n Clustering a data set involving mixed attributes is
n One common way of dealing with mixed
1.
2.
n E.g., if most attributes in a data set are
q we convert ordinal attributes and ratio-scaled
q it is also appropriate to treat symmetric binary
n It does not make much sense to convert a
q but it is frequently done in practice by assigning
n Alternatively, a nominal attribute can be
n This approach computes individual attribute distances and then
n A combination formula, proposed by Gower, is
q The distance dist(xi,xj) is between 0 and 1 q r is the number of attributes q q dij
f is the distance contributed by attribute f, in the range [0,1]
!ij
f =
1 if xif and x jf are not missing 0 if xif or x jf is missing 0 if attribute f is asymmetric and xif and x jf are both 0 ! " # # $ # #
= =
=
r f f ij f ij r f f ij j i
d dist
1 1
) , ( δ δ x x
n If f is a binary or nominal attribute
q distance (4) reduces to
n
equation (3)-lect 10 if all attributes are nominal
n
the simple matching distance (1)-lect 10 if all attributes are symmetric binary
n
the Jaccard distance (2)-lect 10 if all attributes are asymmetric n If f is interval-scaled
q Rf is the value range of f q If all the attributes are interval-scaled, distance (4) reduces to
Manhattan distance
n
Assuming that all attributes values are standardized n Ordinal and ratio-scaled attributes are converted to
f =
f = xif ! x jf
n Basic concepts n K-means algorithm n Representation of clusters n Hierarchical clustering n Distance functions n Data standardization n Handling mixed attributes n Which clustering algorithm to use? n Cluster evaluation n Summary
n Clustering research has a long history
q A vast collection of algorithms are available q We only introduced several main algorithms
n Choosing the “best” algorithm is challenging
q Every algorithm has limitations and works well with certain
q It is very hard, if not impossible, to know what distribution
n
The data may not fully follow any “ideal” structure or distribution required by the algorithms
q One also needs to decide how to standardize the data, to
n Due to these complexities, the common practice is to
1.
2.
n The interpretation of the results must be based on
q insight into the meaning of the original data q knowledge of the algorithms used
n Clustering is highly application dependent and to
n Basic concepts n K-means algorithm n Representation of clusters n Hierarchical clustering n Distance functions n Data standardization n Handling mixed attributes n Which clustering algorithm to use? n Cluster evaluation n Summary
n The quality of a clustering is very hard to evaluate
q We do not know the correct clusters
n Some methods are used
q User inspection n
A panel of experts inspects the resulting clusters and scores them
q Study centroids as spreads q Examine rules (e.g., from a decision tree) that describe the clusters q For text documents, one can inspect by reading n
The final score is the average of the individual scoring
n
Manual inspection is labor intensive and time consuming
n We use some labeled data (for classification)
q Assumption: Each class is a cluster
n Let the classes in the data D be C=(c1, c2,…,ck)
q The clustering method produces k clusters, which
n After clustering, a confusion matrix is constructed
q From the matrix, we compute various measurements:
n For each cluster, we can measure the entropy as
q Pri(cj): proportion of class cj in cluster Di
n The entropy of the whole clustering is
q |Di|/|D| is the weight of cluster Di, proportional to its
j=1 k
i=1 k
n Measures the extent a cluster contains only one class of
n The purity of the whole clustering is
q |Di|/|D| is the weight of cluster Di, proportional to its size
n Precision, recall, and F-measure can be computed as well
q Based on the class that is most frequent in the cluster
i=1 k
n
We can use the total entropy or purity to compare
q
different clustering results from the same algorithm
q
different algorithms
n
Precision, recall and F-measure can be computed as well for each cluster
q
The precision of Science in cluster 1 is 0.89, the recall is 0.83, the F-measure is thus 0.86
n Commonly used to compare different clustering
n A real-life data set for clustering has no class labels
q Thus although an algorithm may perform very well on some
n The fact that it performs well on some label data
n This evaluation method is said to be based on
n Intra-cluster cohesion (compactness):
q Cohesion measures how near the data points in a
q Sum of squared error (SSE) is a commonly used
n Inter-cluster separation (isolation):
q Separation means that different cluster centroids
n In most applications, expert judgments are
n In some applications, clustering is not the primary
n We can use the performance on the primary task to
n For instance, in an application, the primary task is to
q If we can cluster shoppers according to their features, we
q We can evaluate different clustering algorithms based on
q Here, we assume that the recommendation can be reliably
n Basic concepts n K-means algorithm n Representation of clusters n Hierarchical clustering n Distance functions n Data standardization n Handling mixed attributes n Which clustering algorithm to use? n Cluster evaluation n Summary
n Clustering is has along history and still active
q There are a huge number of clustering algorithms q More are still coming every year
n We only introduced several main algorithms. There are
q density based algorithm, sub-space clustering, scale-up methods,
neural networks based methods, fuzzy clustering, co-clustering, etc.
n Clustering is hard to evaluate, but very useful in practice
q This partially explains why there are still a large number of clustering
algorithms being devised every year
n Clustering is highly application dependent and to some
These slides are an adaptation of slides drawn by Tom Mitchell and modified by Liviu Ciortuz
Supervised learning is the simplest and most
How can an agent learn behaviors when it doesn’t
The agent has a task to perform It takes some actions in the world At some later point, it gets feedback telling it how well it
did on performing the task
The agent performs the same task over and over again
This problem is called reinforcement learning:
The agent gets positive reward for tasks done well The agent gets negative reward for tasks done poorly
The goal is to get the agent to act in the world
The agent has to figure out what it did that
This is known as the credit assignment problem
Reinforcement learning can be used to train
playing board games job shop scheduling controlling robot flight/taxy scheduling …
Task: Control learning
make an autonomous agent (robot) to perform actions,
The Q learning algorithm
acquire optimal control strategies from delayed rewards,
even when the agent has no prior knowledge of the effect
Reinforcement Learning is related to dynamic
While DP assumes that the agent/program knows the
effect (and rewards) of all its actions, in RL the agent has to experiment in the real world
Example:
immediate reward
Target function to
Goal: maximize
0 +!r 1 +! 2r 2 +...