machine learning

Question 1

what is accuracy?

Answer 1

correct prediction/ total

Question 2

what is precision?

Answer 2

TP/(TP+FP)

Question 3

what is recall?

Answer 3

TP/TP+FP

Question 4

what is ROC AUC?

Answer 4

ROC curve: receiver operating characteristic curve - plot of TPR (yaxis) and FPR (xaxis) help decide best threshold
AUC: area under the curve higher better help see which categorisation is better

Question 5

what is entropy?

Answer 5

sum(pxlog(1/px))

Question 6

What does linear regression solve?

Answer 6

Used for regression problems

Question 7

How does linear regression work?

Answer 7

It works by fitting a linear equation to observed data. The steps to perform linear regression are:
- First, the sum of squared residuals is calculated.
- Then, this sum is minimized to find the best fit line.

Question 8

what are the parameters of linear regression?

Answer 8

• fit_intercept: Whether to calculate the intercept for this model. If set to False, no intercept will be used in calculations.
• normalize: This parameter is ignored when - - fit_intercept is set to False. If True, the regressors X will be normalized before regression by subtracting the mean and dividing by the l2-norm.

Question 9

What does Naive Bayes solve?

Answer 9

classification problems

Question 10

How does Naive Bayes work?

Answer 10

It works based on Bayes’ theorem with the assumption of independence between every pair of features. Naive Bayes classifiers work well in many real-world situations such as document classification and spam filtering.

Question 11

What are the parameters of NB?

Answer 11

• priors: Prior probabilities of the classes. If specified the priors are not adjusted according to the data.
• var_smoothing: Portion of the largest variance of all features that is added to variances for calculation stability.

Question 12

What does SVM solve?

Answer 12

regression and classification problems

Question 13

How does SVM work?

Answer 13

It works by finding a hyperplane in an N-dimensional space(N — the number of features) that distinctly classifies the data points.

Question 14

What are the parameters in SVM?

Answer 14

• C: Regularization parameter. The strength of the regularization is inversely proportional to C. Must be strictly positive. Larger C –> narrower margin as it penalise wrong classification more.
  - kernel: Specifies the kernel type to be used in the algorithm. It could be ‘linear’, ‘poly’, ‘rbf’, ‘sigmoid’, ‘precomputed’ or a callable.
  - degree: Degree of the polynomial kernel function (‘poly’). Ignored by all other kernels.
  - gamma: Kernel coefficient for ‘rbf’, ‘poly’ and ‘sigmoid’.

Question 15

What is Logistic Regression used for?

Answer 15

binary classification problems

Question 16

How does Logistic regression work?

Answer 16

It works by using a logistic function to model a binary dependent variable.

Question 17

What are the parameters of Logistic Regression?

Answer 17

• penalty: Used to specify the norm used in the penalization. The ‘newton-cg’, ‘sag’ and ‘lbfgs’ solvers support only l2 penalties. ‘elasticnet’ is only supported by the ‘saga’ solver.
  - C: Inverse of regularization strength; must be a positive float. Like in support vector machines, smaller values specify stronger regularization.
  - fit_intercept: Specifies if a constant (a.k.a. bias or intercept) should be added to the decision function.

Question 18

What is K-Means used for?

Answer 18

clustering problems

Question 19

How does K-Means work?

Answer 19

It works by partitioning n observations into k clusters in which each observation belongs to the cluster with the nearest mean, serving as a prototype of the cluster.

Question 20

What are the parameters of K-Means?

Answer 20

• n_clusters: The number of clusters to form as well as the number of centroids to generate.
  - init: Method for initialization, defaults to ‘k-means++’.
  - n_init: Number of time the k-means algorithm will be run with different centroid seeds.

Question 21

What is DBSCAN used for?

Answer 21

clustering problems

Question 22

What is the mechanism of DBSCAN?

Answer 22

It works by defining a cluster as a maximal set of density-connected points. It discovers clusters of arbitrary shape in spatial databases with noise.

Question 23

What are the parameters of DBSCAN?

Answer 23

• eps: The maximum distance between two samples for one to be considered as in the neighborhood of the other.
  - min_samples: The number of samples (or total weight) in a neighborhood for a point to be considered as a core point.

Question 24

What is baggin and boosting used for?

Answer 24

both regression and classification problems

Question 25

What is mechanism of bagging and boosting?

Answer 25

Bagging works by creating subsets of the original dataset, fitting a model to each subset, and then combining the predictions. Boosting works by fitting a model to the data, then fitting additional models to the residuals of the initial model, and then combining the predictions.

Question 26

What is random forest used for?

Answer 26

Used for both regression and classification problems.

Question 27

How does Random Forest work?

Answer 27

It works by creating a multitude of decision trees at training time and outputting the class that is the mode of the classes (classification) or mean prediction (regression) of the individual trees.

Question 28

What are parameters of random forest?

Answer 28

• n_estimators: The number of trees in the forest.
  - max_features: The number of features to consider when looking for the best split.
  - max_depth: The maximum depth of the tree.

Question 29

What is decision trees used for?

Answer 29

Regresssion and classification problems

Question 30

how does decision trees work?

Answer 30

Decision trees split the data into multiple sets. These splits are based on certain conditions, which is why decision trees are also known as Classification and Regression Trees (CART). This process continues on each derived subset in a recursive manner called recursive partitioning.

Question 31

What are the parameters of decision trees?

Answer 31

max_depth: The maximum depth of the tree. This parameter controls over-fitting. Too high value can lead to overfitting and too low value can lead to underfitting.
min_samples_split: The minimum number of samples required to split an internal node. This parameter prevents overfitting. Higher values prevent a model from learning relations which might be highly specific to the particular sample selected for a tree.
min_samples_leaf: The minimum number of samples required to be at a leaf node. This parameter also prevents overfitting similar to min_samples_split.

Question 32

What is Support vector regression (SVR) used for?

Answer 32

regression problems.

Question 33

How does SVR work?

Answer 33

In SVR, we identify a hyperplane with maximum margin such that the maximum number of data points are within that margin. SVR tries to minimize the error within a certain threshold.

Question 34

What are the parameters of SVR?

Answer 34

C: The Regularization parameter. The strength of the regularization is inversely proportional to C. Must be strictly positive.
kernel: Specifies the kernel type to be used in the algorithm. It could be ‘linear’, ‘poly’, ‘rbf’, ‘sigmoid’, ‘precomputed’ or a callable.
degree: Degree of the polynomial kernel function (‘poly’). Ignored by all other kernels.
gamma: Kernel coefficient for ‘rbf’, ‘poly’ and ‘sigmoid’.