5 Data Science Interview Questions
1. What is data science?
Data science is a field that combines the use of “methods, processes, algorithms and systems” to gain knowledge/insights about the data.
2. What is the difference between supervised and unsupervised learning?
In Supervised learning, the algorithm learns from labeled data, meaning that the user has input variables as well as the expected output variable.
In unsupervised learning, the algorithm learns from unlabeled data. Essentially, the goal is for the algorithm to extract patterns on its own. The dataset contains the input variables, but it does not have a specified output variable.
3. How do you validate a machine learning model?
The method used to evaluate a machine learning model depends on the type of data that the user is dealing with.
- Train/test split: this method consists in randomly selecting a subset of your data for training, for instance 80%, and using the other 20% for testing.
- K-fold Cross-Validation: this approach is good to minimize sampling bias. It consists in splitting the training data into k-folds (the user chooses the number of folds), training the data into k-1 folds and then testing the data into the 1 fold that was left out. This is illustrated in the image below.
When it comes to time series data, k-fold cross-validation can also be used, however, the user just needs to make sure that the training data takes precedence over the test data as illustrated in the image below.
4. What is standard deviation?
Standard deviation is a statistical measure of the amount of dispersion or variation in a dataset.
5. What is deep learning?
Deep learning is a field of machine learning inspired by the neurons of the human brain with the goal of processing data for object detection, speech recognition, language translations and many other problems. Essentially, it is a field of machine learning that teaches the computer to learn by example.
References: