2
$\begingroup$

I’ve a dataset about a machine that produces a product. Salient features about the dataset:

• For each data point, some of the features are machine settings (calibration factor, grind time, pressure, water temperature) used to make the product, while few others are stochastic and external in nature, such as water hardness.

• Output says whether the product produced was of good or bad quality.

Goal:

• When I’ve a new data point which says that the product was of bad quality, I want to feed in the corresponding features(machine settings + stochastic external features) to an ML model and predict machine settings which will lead to good quality product. Idea is to have an ML based smart controller which would adjust the machine settings to reduce faulty products being produced.

Which ML models would suit this purpose?

$\endgroup$

closed as too broad by Siong Thye Goh, Mark.F, Sean Owen Feb 3 at 5:12

Please edit the question to limit it to a specific problem with enough detail to identify an adequate answer. Avoid asking multiple distinct questions at once. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • $\begingroup$ Have added more details to the question. $\endgroup$ – Arjun Feb 6 at 0:09
  • $\begingroup$ One approach i can see is to try to predict future values of external factors and also use past data to assess best configuration for the predicted external factors values, us them in order to best respond to the prediction. Algorithms used depend on the data types. You should first analyze external factors and find a way to predict future values. $\endgroup$ – Nemanja Boskovic Feb 6 at 21:03
2
$\begingroup$

Can you explain the problem more precisely? What are inputs (besides control settings), is output same if the inputs are fixed? Are there any stochastic variables(external factors prone to probabilistic behaviour), what is the desired outcome?( Minimizing faulty outcome by choosing a setting or something similar)

$\endgroup$
  • $\begingroup$ Nemanja Boskovic: I've added few more details to my question. Please let me know whether that helps. $\endgroup$ – Arjun Feb 6 at 0:11
1
$\begingroup$

Given your dataset, you can predict whether a given product made using certain machine settings is of bad quality or good quality. But, what you want to achieve (predicting machine settings) is difficult to achieve using a direct machine learning technique as you'll have only one feature and multiple target variables.

However, you can use a knn type approach here. Let us call the data point with bad product for which you want to change settings A. Using knn methodology look for 1(or k) nearest neighbors of A which have good product. Then you can use the machine settings of these neighbors or their average. Though, this is not machine learning!

$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.