Skip to content

HighRadius-Technologies

Directory actions

More options

Directory actions

More options

Latest commit

 

History

History

HighRadius-Technologies

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
README.md
README.md
 
 

README.md

Interview questions (Personal Experience):

1. Confusion matrix for any classification algorithm is showed and what are precision and recall values from the matrix?

  • What is precision: Number of True Positives / (Total number of Total Positives and False Positives)
  • What is Recall : Number of True Positives / (Total Number of Total Positives and False Negatives)

2. How do you deal with overfitting in your machine learning model?

  • Regularization - (LASSO or Ridge Regression) (L1 and L2 regularization).
  • K-Fold cross validation with variable K.
  • Resampling of Train and Test splits of a datasets, sometimes involve out-of-time validation dataset.
  • Dimensionality Reduction in case of many features in a dataset.
  • Ensemble Learning.

3. How do you explain p-value to a layman? - Hypothesis testing.

  • p-value is metric by which we decide statistically significant variables.
  • It's a measure of how extreme an observed value is under the assumed null hypothesis: the smaller it is, the more extreme the observation. We can define p-value as the smallest significance level at which the null hypothesis would be rejected.
  • As the p-value gets smaller, we start wondering if the null hypothesis really is true and well maybe we should change our minds (and reject the null hypothesis).

4. How do you deal with concurrent predictions from decision Trees in ensemble algorithms like random forest?

  • Do some google research on this, there are multiple techniques to deal with this.