Naïve Bayes:Naïve bayes is simple and most commonly used classifier. It usesall the features in feature vector. Naïve bayes analyzes the feature vectorindividually as they are equally independent to each other. It is a probabilistic classifier, the conditionalprobability of naïve bayes is defined as P(X|yj) =_mi=1P(xijyj) (1)Here ‘X’ is feature vector, it is defined as X={x/,x2,…….xm}. Yj is the class label.

When a labelled is to be classified it splits the data into single word features. Naïve bayes use probabilities , which iscomputed in training stage to calculate the conditional probability of combined features in order to predict class. Themain advantage of Naïve Bayes classifier is it makes use of all the evidencewhich is available in order to make text classificationSupportVector Machine: The supportvector machine classifier is considered to be non-probabilistic binary linearclassifier. Support vector machine is a supervised model.

This classifier useslarge margins for the classification. It separates the data using a hyperplane. Support vector machine uses the concept of decision planes that definesdecision boundaries.g(X)=wr (X)+b (2) ‘X’ is afeature vector, ‘w’ is a weight vector and ‘b’ is bias vector . O is non linearmapping from input space to high dimensional features space.

Here ‘w’ and ‘b’both are learned automatically on training set . SVM can be used for thepattern recognition.Decisiontree: The decision tree classifier is a supervised learningalgorithm which can used for both classification and regression tasks. It can be adapted almost to any type of data.

Itdivides the training data into small parts in order to identify patterns sothat they can be used for classification. This algorithm is specifically usedwhere there are many hierarchical categorical distinctions can be madepriyanka.The structure of a decision tree consists of rootnode, decision node and leaf node.

Theroot node represents the entire data set and decision node performs computationand leaf node produces the classification. It is particularly used when thereare many hierarchical categorical distinctions can be made. In training phase,this algorithm learns what are the decisions have to be made in order to splitthe labelled data into its classes. Passing the data through tree, a unknowninstance is classified. The computation which takes place in each of thedecision node usually compares the selected feature with predeterminedconstant, the decision will be made based on whether the feature is greater orless than the constant by creating two way split in the tree.

The data will beeventually passed through these decision nodes until that reaches a leaf nodewhich represent its assigned class.MaximumEntropy: Maximum entropy provides a machine learning technique forprediction. This is also known as multinomial logistic model. The maximumentropy maximizes the entropy that is defined on the conditional probability distribution.It can even handles the overlap feature and is same as the logistic regression whichfinds the distribution over classes.

It also follows some certain feature constraintsgeetika.’C’is the class, ‘d’ is the sentence, ? is the weight vectorwhereas, the weight vector calculates the importance of a feature. More weightdenotes that a particular is strong enough in class.Random Forests: Randomforests are an ensemble learning method for classification that performs withthe help of building a large number of selection trees at training time andoutputting the class that is mode for the training output through individualtrees. It produces multi-altitude selection trees at input phase and output isgenerated inside within the form of multiple decision trees. The correlationamong trees is reduced by randomly deciding on trees and as a result theprediction strength will increase and resultsto boom in performance.

ref random predictions are made by aggregating thepredictions of various ensemble information units.