8

First of all a picture should not be taken to define if there are or not groups on your data, since no matter what projection you use (linear with PCA or manifold with tSNE) you are reducing a 64-dimensional space into a 2-dimensional space, that's a lot of information lost. Secondly, as far as I know there is no theorem that guarantees you can find clusters ...


6

This is true of any data analytics endeavor. You don't have ANY guarantees that you're going to find what you are looking for in your data. You have a theory, question, assumptions... and you collect data to see if that fits the reality. Careful though, absence of prove is not prove of absence. There are reasons why you might not see what you expect. In your ...


2

64 features (like, total watching time, percent of ads skipped, movies vs. shows, etc.). All variables are either numerical or binary. There are several problems with this. It's best to see them when trying to decompose your data into so called main-effects, at least as a mental experiment or excercise. Don't see what your objective function y(par1, par2, ....


1

There are several options: Hand-picked rules - Given domain expertise, manually choose the threshold values to create the four clusters. Machine learning - Set the number of clusters to four. Then use any clustering algorithm (e.g., k-means, Gaussian mixture model, DBSCAN, spectral). This has the advantage of learning the threshold values. Choosing the ...


1

So, Clustering is "Unsupervised" learning : You make groups in which elements look like each-other. In Unsupervised learning, you don't have a Label that you look for. Here, your problem is to Classify text between 3 categories : Sports, Foreign, Local. Those 3 categories ARE labels : You know you have news about those 3 subjects, and want to make ...


1

The most intuitive way of visualizing your cluster results would be by using a linear projection like PCA. In this way you can visualize for example the first 3 components and assign a color to each point according to cluster_id Also important, you should in this case check the explained_variance as measure of how reliable the projection is, since you are ...


1

The standard evaluation would be to count the proportion of correct predictions. The most basic version would be to count 3 instances for every event: where, when, what. For example if the three questions are answered correctly the score for this event is 3/3. Note that the case where one of the questions has no answer should be counted normally, i.e. if ...


1

One thing you can use is the cluster_centers_ (it's an attribute from your trained KMeans). It'll give you a numpy ndarray with the position, according to each variable, of the centroid of each cluster. You can compare it to your mean value for each variable (or, if normalized, to zero), so you have an idea. For example, if you have a variable Age, with a ...


Only top voted, non community-wiki answers of a minimum length are eligible