# Classification of data tables (each table is an item)

I have to work on a binary classification task where single items to be classified are not single rows of a data matrix, but groups of rows. In other words, I have $$N$$ data tables of varying size $$n_i \times p$$ for $$1 \le i \le N$$, and I have to train a classification algorithm over those tables, where the target labels are $$N$$ in number, and of course refer to the tables.

Actually, I have already thought about some strategies to address the problem. Those are not the point of my question, so I will only dwell on the simplest one, so I can give you a clearer idea of my situation: I could simply take the mean for each of the $$p$$ columns, for each of the $$N$$ groups, and then train the algorithm, because at that point I would have a simple $$N \times p$$ data table: one observation, one class. Since the means alone seem too little, I could also take the variances and covariances, so to get a training set of size $$N \times (p + \frac{p(p+1)}{2})$$.

Anyway, my question is about literature: I can't find any paper about this kind of problem, not even one. That's probably because I couldn't look in the right places, because this doesn't appear to me as something so strange and unusual.

I want to know if this kind of problem has a name, that I ignore, and I would also like to be addressed to the scientific work that has been published about it. The more, the better.

Edit: I found this related question, where the first answer points to a python package that automatically extracts features from tables related to the main dataset. That package is cited in a few papers where the problem I expose is not really considered according to my definition. It seems to me that we are just starting to figure how we can exploit such amounts of data.

• Do you mean you have $N$ data points with $N$ different labels? If so, I suppose one sensible approach would be to use nearest neighbour. Just classify any unobserved (new) table by seeing which of the existing $N$ tables it is closest to. – Tim Mak May 8 '20 at 7:43
• well I wouldn't call them points since they are not vectors, but yes, I thought about that, even if missings are quite a problem in my case, when it comes to compute distances. – carlo May 8 '20 at 9:30
• From the data analyst perspective, this is an aggregation problem: the shape of your data isn't ideal, so how can you meaningfully summarize the data into a better shape? Tim Mak's idea is a good one, averaging or taking a median is another. There are many, many ways to aggregate data; you need to choose one that makes sense for your application. – Adrian Keister May 8 '20 at 14:23
• Since you mentioned you have $N$ data tables of varying size, can you tell us more about why you have these additional rows per data point? Is it so that you have repeated measurements for the observations, and number of repetitions varied during data collections? How different are these rows in individual observation tables? – PAF May 11 '20 at 21:56
• @PAF right, I have $n_i$ independent observations related to $i$-th phenomenon. – carlo May 12 '20 at 11:17