Which machine learning methods are best for predicting?

Question

Neural networks

Given: A database with information about all the failures that have occurred in production.
Task: Develop a system for predicting future failures.
What methods are more suitable? Which entry threshold is not very high?

Answer 1

1. Analysis of time series.
2. Building regression models.
3. Classification (by all known methods - from metric and support vector systems to the use of decision trees, do not forget about the use of ensembles of methods)
4. Clustering and principal component analysis.
5. All the same but based on the use of artificial neural networks.
All fit. You just need to understand what, when and how.
With a threshold - well, it depends on your entry level.