C
C
chaschev2020-11-03 01:05:26
Machine learning
chaschev, 2020-11-03 01:05:26

Simple machine learning API?

Has a simple machine learning API for housewives appeared so that it can be used effortlessly in normal code - for example:

Example 1. Login: user clicks on menu items, array (user, menuItem, time). Output:
shortcut menu for a given user to menu items

It is clear that it is possible to write a solution that performs the same task, but why not do it in automatic mode?

Example 2. There are indications of the cost of electricity. Need to predict the best time to do laundry based on activity data from my sports bracelet.

Example 3. There is a vector of economic data of companies. P/E, EPS, Revenue, EBIDTA, P/B etc. It is necessary to find the most profitable / calculate one value through another. For example, calculate EPS based on the historical values ​​of the debt to equity ratio and the growth dynamics of the sector

Answer the question

In order to leave comments, you need to log in

3 answer(s)
X
xmoonlight, 2020-11-03
@xmoonlight

The minimum is a task-based decision tree for selecting the desired type of NN.
And then - the configuration of the selected type of NS, based on the input data, to obtain the requested result.
It's not so simple... But soon.

M
mayton2019, 2020-11-03
@mayton2019

Not everything that is considered math statistics should be called machine learning. The first two points do not require ML at all.
The third one is so complicated and incomprehensible in its formulation that it just needs to be discussed for a very long time. What's in and what's out.
So that it would not be like in a joke about "navigator device 100" ...

D
dmshar, 2020-11-03
@dmshar

If example 1 can still be attributed to issues that may attract the attention of the creators of the corresponding software, then example 2 is somehow incomprehensibly formulated (the best time for washing is when the cost of electricity is minimal, i.e. at night - what is there to analyze from your bracelet?) and example 3 is just an algorithmic problem. And if we talk about the "prediction" of the values ​​of the time series of the corresponding indications, then it is also well developed and is solved by the appropriate methods from the machine learning libraries.

Didn't find what you were looking for?

Ask your question

Ask a Question

731 491 924 answers to any question