Iso Cluster Unsupervised Classification - Combining Classes

[JoeCook4]

Hi guys,

Let me preface my question by saying that I have very little experience with ArcGIS. Most of what I know, I've had to learn through articles on the web as well as the support here on ESRI's website. With that said, I am trying to combine classes after just running an ISODATA Cluster Unsupervised Classification. My final product needs to have around 5-10 classes. From what I have read, I am going to need to use the Swipe, Flicker and Identify tools to discover agreement (or disagreement) between points falling in the same class. I understand how the Identify tool works but how do the other two work and how should I use them to help me recombine classes? Also, how should I be using the Reclassify tool?

One more question, when I run the ISODATA Cluster Unsupervised Classification, I ask for 16 classes yet it only gives me 3-4. Why does it do this? How can I get more classes?

ANY help would be GREATLY appreciated! Thanks guys and sorry for my noobiness.

BattlePope

[JeffreySwain]

While you have requested the larger number of classes, there are reasons summarized in the help that indicate why you would have less.

The specified Number of classes value is the maximum number of clusters that can result from the clustering process. However, the number of clusters in the output signature file may not be the same as the number specified for the number of classes. This situation occurs in the following cases:

    The values of data and the initial cluster means are not evenly distributed. In certain ranges of cell values, the frequency of occurrences for these clusters may be next to none. Consequently, some of the originally predefined cluster means may not have a chance to absorb enough cell members.
    Clusters consisting of fewer cells than the specified Minimum class size value will be eliminated at the end of the iterations.
    Clusters merge with neighboring clusters when the statistical values are similar after the clusters become stable. Some clusters may be so close to each other and have such similar statistics that keeping them apart would be an unnecessary division of the data.

The help documentation cited is here, for the iso cluster.  I would consider the data and then take a look at the settings specified in the processing, perhaps altering the minimum size would prevent the data from being lumped together.