algorithms

what is supervised learning

machine learning task of inferring a function from labeled training data

what is machine learning

machine learning explores the study and construction of algorithm that can learn from and make predictions on data

give examples of supervised learning algorithms

support vector machines, regression, naive bayes, decision trees,

what is unsupervised learning

type of machine learning algorithm used to draw interferences from datasets consisting of input data without labeled responses

give example of unsupervised learning algorithms

clustering, anomaly detection, k-means for clustering

what are the various classification algorithms

decision trees, svm, logistics regression, naive bayes

what is logistics regression?

Is a technique to predict the binary outcome from linear combination of predictor variables

what is linear regression?

statistical technique where the score of a variable Y is predicted from the score of second variable X. X is referred to as the predictor variable and Y as criterion variable

explain svm algorithm

supervised ml. both regression and class. svm uses hyper planes to separate out different classes based on the provided kernel function

what are the different kernels functions in svm

linear, polynomial, radial basis, sigmoid

what is random forest

each tree gives a classification. the forest chooses the classification having the most votes. in regression it takes average

explain decision tree

For regression and classification. it breaks down a data set for a smaller subsets while at the same time an associated decision tree is incrementally developed. the final result is a tree with nodes and leafs

boosting

an iterative technique which adjust the weight of an observation based on last classification. if an observation was classified incorrectly, it tries to increase the weight of this observation and vice versa

what is confusion matrix

2x2 Table contains 4 outputs provided by binary classifier. various measures, such as error-rate, accuracy, specificit, sensitivity, precision and recall

precision measure

TP/(TP+FP) precision is a good to determine, when the costs of FP is high.

recall measure

TP/(TP+FN) recall shall be the model metric when there is a high cos associated with False Negative

F1 score

2x(precision x recall) /(precision + recall) might be a better measure to use if we need to seek a balance between Precision and Recall AND there is an uneven class distribution(large number of Actual Negatives)

what is data sampling?

statistical analysis of data. it is used to select, manipulate, and examine a representative subgroup of data points that allow you to identify trends

what is selection bias

unrepresentative sample of data. it is when the data that has been mined, cleaned, and prepared for modeling is not illustrative of the data that the model will see once it is in use

what is regularization?

adds penalty to a model as complexity increases. this prevents overfitting.

what is bias?

bias is error introduced in your model due to over simplification of ML algorithm. it can lead to under fitting

Variance

Variance is error introduced in your model due to complex ML algorithm, your model learns noise also from the training data set and performs bad on test data set. it can lead high sensitivity and overfitting

Gradient

Gradient is the direction and magnitude calculated during training of a neural network that is used to update the network weights in the right direction and by the right amount

explain how ROC curve works

the roc is a graphical representation of the contrast between TP rates and FP rates at various thresholds.