Many algorithms for clustering are available. A popular algorithm is the K-means where, based on a given number of clusters, the algorithm iterates to find best clusters for the objects.
What method do you use to determine the number of clusters in the data in k-means clustering?
Does any package available in R contain the V-fold cross-validation method for determining the right number of clusters?
Another well used approach is Expectation Maximization (EM) algorithm which assigns a probability distribution to each instance which indicates the probability of it belonging to each of the clusters.
Is this algorithm implemented in R?
If it is, does it have the option to automatically select the optimum number of clusters by cross validation?
Do you prefer some other clustering method instead?
For large "sparse" datasets i would seriously recommend "Affinity propagation" method.
It has superior performance compared to k means and it is deterministic in nature.
http://www.psi.toronto.edu/affinitypropagation/
It was published in journal "Science".
However the choice of optimal clustering algorithm depends on the data set under consideration. K Means is a text book method and it is very likely that some one has developed a better algorithm more suitable for your type of dataset/
This is a good tutorial by Prof. Andrew Moore (CMU, Google) on K Means and Hierarchical Clustering.
http://www.autonlab.org/tutorials/kmeans.html
Last week I coded up such an estimate-the-number-of-clusters algorithm for a K-Means clustering program. I used the method outlined in:
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.70.9687&rep=rep1&type=pdf
My biggest implementation problem was that I had to find a suitable Cluster Validation Index (ie error metric) that would work. Now it is a matter of processing speed, but the results currently look reasonable.
Related
I could use some advice on methods in R to determine the optimal number of clusters and later on describe the clusters with different statistical criteria. I’m new to R with basic knowledge about the statistical foundations of cluster analysis.
Methods to determine the number of clusters: In the literature one common method to do so is the so called "Elbow-criterion" which compares the Sum of Squared Differences (SSD) for different cluster solutions. Therefore the SSD is plotted against the numbers of Cluster in the analysis and an optimal number of clusters is determined by identifying the “elbow” in the plot (e.g. here: https://en.wikipedia.org/wiki/File:DataClustering_ElbowCriterion.JPG)
This method is a first approach to get a subjective impression. Therefore I’d like to implement it in R. The information on the internet on this is sparse. There is one good example here: http://www.mattpeeples.net/kmeans.html where the author also did an interesting iterative approach to see if the elbow is somehow stable after several repetitions of the clustering process (nevertheless it is for partitioning cluster methods not for hierarchical).
Other methods in Literature comprise the so called “stopping rules”. MILLIGAN & COOPER compared 30 of these stopping rules in their paper “An examination of procedures for determining the number of clusters in a data set” (available here: http://link.springer.com/article/10.1007%2FBF02294245) finding that the Stopping Rule from Calinski and Harabasz provided the best results in a Monte Carlo evaluation. Information on implementing this in R is even sparser.
So if anyone has ever implemented this or another Stopping rule (or other method) some advice would be very helpful.
Statistically describe the clusters:For describing the clusters I thought of using the mean and some sort of Variance Criterion. My data is on agricultural land-use and shows the production numbers of different crops per Municipality. My aim is to find similar patterns of land-use in my dataset.
I produced a script for a subset of objects to do a first test-run. It looks like this (explanations on the steps within the script, sources below).
#Clusteranalysis agriculture
#Load data
agriculture <-read.table ("C:\\Users\\etc...", header=T,sep=";")
attach(agriculture)
#Define Dataframe to work with
df<-data.frame(agriculture)
#Define a Subset of objects to first test the script
a<-df[1,]
b<-df[2,]
c<-df[3,]
d<-df[4,]
e<-df[5,]
f<-df[6,]
g<-df[7,]
h<-df[8,]
i<-df[9,]
j<-df[10,]
k<-df[11,]
#Bind the objects
aTOk<-rbind(a,b,c,d,e,f,g,h,i,j,k)
#Calculate euclidian distances including only the columns 4 to 24
dist.euklid<-dist(aTOk[,4:24],method="euclidean",diag=TRUE,upper=FALSE, p=2)
print(dist.euklid)
#Cluster with Ward
cluster.ward<-hclust(dist.euklid,method="ward")
#Plot the dendogramm. define Labels with labels=df$Geocode didn't work
plot(cluster.ward, hang = -0.01, cex = 0.7)
#here are missing methods to determine the optimal number of clusters
#Calculate different solutions with different number of clusters
n.cluster<-sapply(2:5, function(n.cluster)table(cutree(cluster.ward,n.cluster)))
n.cluster
#Show the objects within clusters for the three cluster solution
three.cluster<-cutree(cluster.ward,3)
sapply(unique(three.cluster), function(g)aTOk$Geocode[three.cluster==g])
#Calculate some statistics to describe the clusters
three.cluster.median<-aggregate(aTOk[,4:24],list(three.cluster),median)
three.cluster.median
three.cluster.min<-aggregate(aTOk[,4:24],list(three.cluster),min)
three.cluster.min
three.cluster.max<-aggregate(aTOk[,4:24],list(three.cluster),max)
three.cluster.max
#Summary statistics for one variable
three.cluster.summary<-aggregate(aTOk[,4],list(three.cluster),summary)
three.cluster.summary
detach(agriculture)
Sources:
http://www.r-tutor.com/gpu-computing/clustering/distance-matrix
How to apply a hierarchical or k-means cluster analysis using R?
http://statistics.berkeley.edu/classes/s133/Cluster2a.html
The elbow criterion as your links indicated is for k-means. Also the cluster mean is obviously related to k-means, and is not appropriate for linkage clustering (in particular not for single-linkage, see single-link-effect).
Your question title however mentions hierarchical clustering, and so does your code?
Note that the elbow criterion does not choose the optimal number of clusters. It chooses the optimal number of k-means clusters. If you use a different clustering method, it may need a different number of clusters.
There is no such thing as the objectively best clustering. Thus, there also is no objectively best number of clusters. There is a rule of thumb for k-means that chooses a (maybe best) tradeoff between number of clusters and minimizing the target function (because increasing the number of clusters always can improve the target function); but that is mostly to counter a deficit of k-means. It is by no means objective.
Cluster analysis in itself is not an objective task. A clustering may be mathematically good, but useless. A clustering may score much worse mathematically, but it may provide you insight to your data that cannot be measured mathematically.
This is a very late answer and probably not useful for the asker anymore - but maybe for others. Check out the package NbClust. It contains 26 indices that give you a recommended number of clusters (and you can also choose your type of clustering). You can run it in such a way that you get the results for all the indices and then you can basically go with the number of clusters recommended by most indices. And yes, I think the basic statistics are the best way to describe clusters.
You can also try the R-NN Curves method.
http://rguha.net/writing/pres/rnn.pdf
K means Clustering is highly sensitive to the scale of data e.g. for a person's age and salary, if not normalized, K means would consider salary more important variable for clustering rather than age, which you do not want. So before applying the Clustering Algorithm, it is always a good practice to normalize the scale of data, bring them to the same level and then apply the CA.
I have read a couple of different articles on how PAM selects the initial medoids but I am getting conflicting views.
Some propose that the k first medoids are selected randomly, while others suggest that the algorithm selects initially the k representative medoids in the dataset (not clarifying how that "representativeness" happens though). Below I have listed these resources:
Medoid calculation
Drawbacks of K-Medoid (PAM) Algorithm
https://paginas.fe.up.pt/~ec/files_1112/week_06_Clustering_part_II.pdf
https://www.datanovia.com/en/lessons/k-medoids-in-r-algorithm-and-practical-examples/
1) My question would be if someone could explain in more detail how the algorithm selects the initial k medoids as from what I understand different initial selection can lead to different results.
2) Also is that one of the reasons of using CLARA (apart from minimizing computing time and RAM storage problem) - that is to find medoids through resampling that are the "optimal" options?
I am using R as a parenthesis, with the function pam(). Open to other functions in other libraries if there is a better alternative I am not aware of.
Read the original sources.
There is a lot of nonsense written later, unfortunately.
PAM consists of two algorithms:
BUILD to choose the initial medoids (not randomly)
SWAP to make the best improvements (not k-means style)
The k-means style algorithm works much worse than PAM. Any description of PAM that doesn't mention these two parts is inaccurate (and there are quite some of these...)
The R package seems to use the real PAM algorithm:
By default, when medoids are not specified, the algorithm first looks for a good initial set of medoids (this is called the build phase). Then it finds a local minimum for the objective function, that is, a solution such that there is no single switch of an observation with a medoid that will decrease the objective (this is called the swap phase)
CLARA clearly will find worse solutions than PAM, as it runs PAM on a sample, and I'd the optimum medoids are not in the sample, then they cannot be found.
I could use some advice on methods in R to determine the optimal number of clusters and later on describe the clusters with different statistical criteria. I’m new to R with basic knowledge about the statistical foundations of cluster analysis.
Methods to determine the number of clusters: In the literature one common method to do so is the so called "Elbow-criterion" which compares the Sum of Squared Differences (SSD) for different cluster solutions. Therefore the SSD is plotted against the numbers of Cluster in the analysis and an optimal number of clusters is determined by identifying the “elbow” in the plot (e.g. here: https://en.wikipedia.org/wiki/File:DataClustering_ElbowCriterion.JPG)
This method is a first approach to get a subjective impression. Therefore I’d like to implement it in R. The information on the internet on this is sparse. There is one good example here: http://www.mattpeeples.net/kmeans.html where the author also did an interesting iterative approach to see if the elbow is somehow stable after several repetitions of the clustering process (nevertheless it is for partitioning cluster methods not for hierarchical).
Other methods in Literature comprise the so called “stopping rules”. MILLIGAN & COOPER compared 30 of these stopping rules in their paper “An examination of procedures for determining the number of clusters in a data set” (available here: http://link.springer.com/article/10.1007%2FBF02294245) finding that the Stopping Rule from Calinski and Harabasz provided the best results in a Monte Carlo evaluation. Information on implementing this in R is even sparser.
So if anyone has ever implemented this or another Stopping rule (or other method) some advice would be very helpful.
Statistically describe the clusters:For describing the clusters I thought of using the mean and some sort of Variance Criterion. My data is on agricultural land-use and shows the production numbers of different crops per Municipality. My aim is to find similar patterns of land-use in my dataset.
I produced a script for a subset of objects to do a first test-run. It looks like this (explanations on the steps within the script, sources below).
#Clusteranalysis agriculture
#Load data
agriculture <-read.table ("C:\\Users\\etc...", header=T,sep=";")
attach(agriculture)
#Define Dataframe to work with
df<-data.frame(agriculture)
#Define a Subset of objects to first test the script
a<-df[1,]
b<-df[2,]
c<-df[3,]
d<-df[4,]
e<-df[5,]
f<-df[6,]
g<-df[7,]
h<-df[8,]
i<-df[9,]
j<-df[10,]
k<-df[11,]
#Bind the objects
aTOk<-rbind(a,b,c,d,e,f,g,h,i,j,k)
#Calculate euclidian distances including only the columns 4 to 24
dist.euklid<-dist(aTOk[,4:24],method="euclidean",diag=TRUE,upper=FALSE, p=2)
print(dist.euklid)
#Cluster with Ward
cluster.ward<-hclust(dist.euklid,method="ward")
#Plot the dendogramm. define Labels with labels=df$Geocode didn't work
plot(cluster.ward, hang = -0.01, cex = 0.7)
#here are missing methods to determine the optimal number of clusters
#Calculate different solutions with different number of clusters
n.cluster<-sapply(2:5, function(n.cluster)table(cutree(cluster.ward,n.cluster)))
n.cluster
#Show the objects within clusters for the three cluster solution
three.cluster<-cutree(cluster.ward,3)
sapply(unique(three.cluster), function(g)aTOk$Geocode[three.cluster==g])
#Calculate some statistics to describe the clusters
three.cluster.median<-aggregate(aTOk[,4:24],list(three.cluster),median)
three.cluster.median
three.cluster.min<-aggregate(aTOk[,4:24],list(three.cluster),min)
three.cluster.min
three.cluster.max<-aggregate(aTOk[,4:24],list(three.cluster),max)
three.cluster.max
#Summary statistics for one variable
three.cluster.summary<-aggregate(aTOk[,4],list(three.cluster),summary)
three.cluster.summary
detach(agriculture)
Sources:
http://www.r-tutor.com/gpu-computing/clustering/distance-matrix
How to apply a hierarchical or k-means cluster analysis using R?
http://statistics.berkeley.edu/classes/s133/Cluster2a.html
The elbow criterion as your links indicated is for k-means. Also the cluster mean is obviously related to k-means, and is not appropriate for linkage clustering (in particular not for single-linkage, see single-link-effect).
Your question title however mentions hierarchical clustering, and so does your code?
Note that the elbow criterion does not choose the optimal number of clusters. It chooses the optimal number of k-means clusters. If you use a different clustering method, it may need a different number of clusters.
There is no such thing as the objectively best clustering. Thus, there also is no objectively best number of clusters. There is a rule of thumb for k-means that chooses a (maybe best) tradeoff between number of clusters and minimizing the target function (because increasing the number of clusters always can improve the target function); but that is mostly to counter a deficit of k-means. It is by no means objective.
Cluster analysis in itself is not an objective task. A clustering may be mathematically good, but useless. A clustering may score much worse mathematically, but it may provide you insight to your data that cannot be measured mathematically.
This is a very late answer and probably not useful for the asker anymore - but maybe for others. Check out the package NbClust. It contains 26 indices that give you a recommended number of clusters (and you can also choose your type of clustering). You can run it in such a way that you get the results for all the indices and then you can basically go with the number of clusters recommended by most indices. And yes, I think the basic statistics are the best way to describe clusters.
You can also try the R-NN Curves method.
http://rguha.net/writing/pres/rnn.pdf
K means Clustering is highly sensitive to the scale of data e.g. for a person's age and salary, if not normalized, K means would consider salary more important variable for clustering rather than age, which you do not want. So before applying the Clustering Algorithm, it is always a good practice to normalize the scale of data, bring them to the same level and then apply the CA.
I have ran clv package which consists of S_Dbw and SD validity indexes for clustering purposes in R commander. (http://cran.r-project.org/web/packages/clv/index.html)
I evaluated my clustering results from DBSCAN, K-Means, Kohonen algorithms with S_Dbw index. but for all these three algorithms S_Dbw is "Inf".
Is it "Infinite" meaning? Why did i confront with "Inf". Is there any problem in my clustering results?
In general, when is S_Dbw index result "Inf"?
Be careful when comparing different algorithms with such an index.
The reason is that the index is pretty much an algorithm in itself. One particular clustering will necessarily be the "best" for each index. The main difference between an index and an actual clustering algorithm is that the index doesn't tell you how to find the "best" solution.
Some examples: k-means minimizes the distances from cluster members to cluster centers. Single-link hierarchical clustering will find the partition with the optimal minimum distance between partitions. Well, DBSCAN will find the partitioning of the dataset, where all density-connected points are in the same partition. As such, DBSCAN is optimal - if you use the appropriate measure.
Seriously. Do not assume that because one algorithm scores higher than another in a particular measure means that the algorithm works better. All that you find out this way is that a particular algorithm is more (cor-)related to a particular measure. Think of it as a kind of correlation between the measure and the algorithm, on a conceptual level.
Using a measure for comparing different results of the same algorithm is different. Then obviously there shouldn't be a benefit from one algorithm over itself. There might still be a similar effect with respect to parameters. For example the in-cluster distances in k-means obviously should go down when you increase k.
In fact, many of the measures are not even well-defined on DBSCAN results. Because DBSCAN has the concept of noise points, which the indexes do not AFAIK.
Do not assume that the measure will either give you an indication of what is "true" or "correct". And even less, what is useful or new. Because you should be using cluster analysis not to find a mathematical optimum of a particular measure, but to learn something new and useful about your data. Which probably is not some measure number.
Back to the indices. They usually are totally designed around k-means. From a short look at S_Dbw I have the impression that the moment one "cluster" consists of a single object (e.g. a noise object in DBSCAN), the value will become infinity - aka: undefined. It seems as if the authors of that index did not consider this corner case, but only used it on toy data sets where such situations did not arise. The R implementation can't fix this, without diverting from the original index and instead turning it into yet another index. Handling noise objects and singletons is far from trivial. I have not yet seen an index that doesn't fail in one way or another - typically, a solution such as "all objects are noise" will either score perfect, or every clustering can trivially be improved by putting each noise object to the nearest non-singleton cluster. If you want your algorithm to be able to say "this object doesn't belong to any cluster" then I do not know any appropriate index.
The IEEE floating point standard defines Inf and -Inf as positive and negative infinity respectively. It means your result was too large to represent in the given number of bits.
In R language is there a predict function in clustering like the way we have in classification?
What can we conclude from the clustering graph result that we get from R, other that comparing two clusters?
Clustering does not pay attention to prediction capabilities. It just tries to find objects that seem to be related. That is why there is no "predict" function for clustering results.
However, in many situations, learning classifiers based on the clusters offers an improved performance. For this, you essentially train a classifier to assign the object to the appropriate cluster, then classify it using a classifier trained only on examples from this cluster. When the cluster is pure, you can even skip this second step.
The reason is the following: there may be multiple types that are classified with the same label. Training a classifier on the full data set may be hard, because it will try to learn both clusters at the same time. Splitting the class into two groups, and training a separate classifier for each, can make the task significantly easier.
Many packages offer predict methods for cluster object. One of such examples is clue, with cl_predict.
The best practice when doing this is applying the same rules used to cluster training data. For example, in Kernel K-Means you should compute the kernel distance between your data point and the cluster centers. The minimum determines cluster assignment (see here for example). In Spectral Clustering you should project your data point dissimilarity into the eigenfunctions of the training data, compare the euclidean distance to K-Means centers in that space, and a minimum should determine your cluster assignment (see here for example).