E
E
einleif2016-08-11 11:39:16
data mining
einleif, 2016-08-11 11:39:16

Which machine learning algorithm to choose?

I need to choose a machine learning algorithm to solve one problem: identifying a relationship between individual elements of two different sets of different objects.
Each object (an individual element of a set) is characterized by a set of parameters, and there is a certain relationship between these characteristic parameters, on the basis of which it is possible to draw a conclusion about the relationship between elements of different sets. But there is no exact mathematical model, so I think you can try to use machine learning to find relationships between elements of different sets.
Objects in sets are arranged by index (or by time), there can be no cross-links. This is one of the main criteria for establishing a connection between objects.
I am an absolute beginner in machine learning and everything connected with it, but I am well acquainted with applied programming and feel the strength to try and deal with machine learning. :) But I just got lost in a huge number of different algorithms, methods and principles of machine learning.
What algorithms can suit me for solving such a problem?

Answer the question

In order to leave comments, you need to log in

3 answer(s)
S
Sergey, 2016-08-11
@begemot_sun

> revealing the relationship between the individual elements of two different sets of different objects.
this is not a formalized statement.
When you formalize it, then come :)

A
Alibaba2018, 2019-04-30
@Alibaba2018

A map for beginners on how best to choose algorithms in scikit learn:
https://scikit-learn.org/stable/tutorial/machine_l...
You most likely need to watch either linear regression or KMeans Clusterin IMHO,
depending on the number of parameters

V
Vlad_Fedorenko, 2016-08-11
@Vlad_Fedorenko

The question here is not which algorithm to use, but how to present the data so that you can set your favorite logistic regression or gradient boosting.
If there are not very many objects, then you can generate all pairs (a, b), where a is from the first set, and b is from the second, come up with signs for these pairs and train the algorithm on "there is a connection / no connection"

Didn't find what you were looking for?

Ask your question

Ask a Question

731 491 924 answers to any question