DecisiontreeDecision tree methodologyis a usually used data mining method for starting classification systems basedon multiple covariates or for developing forecast algorithms for a targetvariable.The basic concept of thedecision tree 1. Nodes. Thereare three types of nodes. (Lu and Song, 2017)- A root hub, additionally called a choicehub, symbolizes a decision that will bring about the segment of all recordsinto at least two similarly selective subsets. – Internal hubs, additionally called shothubs, symbolize one of the conceivable choices accessible at that reality inthe tree structure, the upper edge of the hub is associated with its parent huband the most profound edge is associated with its kid hubs or leaf hubs.
– Leaf hubs, likewise called end hubs, speakto the last impact of a blend of choices or occasions.2. Branches. (Lu and Song, 2017)- Branches symbolize chance outcomes or events that originate from roothubs and inward hubs.
– A decision tree demonstrate is composed utilizing a pecking order ofbranches. Every way from the root hub over inner hubs to a leaf hub speaks to agrouping choice run the show. – These decision tree ways can likewise be spoken to as ‘assuming at thatpoint’ rules.3. Splitting.
(Lu and Song, 2017)- Only the inputvariables interrelated to the target variable are charity to split parent nodesinto purer child nodes of the target variable. – Both separate inputvariables and incessant input variables which are collapsed into two or morecategories can be used. – When building themodel one need first identify the most important input variables, and thensplit records at the root node and at succeeding internal nodes into two ormore classes or ‘bins’ based on the status of these variables. The type of the decision tree · Classification tree analysis is when the forecastoutcome is the class to which the data belongs.· Regression tree analysis is when thepredicted outcome can be considered a real number (e.
g. the price of a house,or a patient’s length of stay in a hospital). Decision tree can quickly express complex optionsplainly. Furthermore, can without much of a spring adjust a decision tree asnew data storms up noticeably available. Set up a decision tree to look at how shiftinginformation regards influence different choice options. Standard decision tree certificationis anything but difficult to receive. You can think about contending choiceseven without finish data as far as threat and likely esteem.
(Anon, 2017) 2. Logistic Regression – Logistic regression is used to find theprobability of event=Success and event=Failure. We should use logisticregression when the dependent variable is binary (0/ 1, True/ False, Yes/ No)in nature. – The binarylogistic model is charity to estimate the probability of a binary responsebased on one or more predictor (or independent) variables (features). – It allowsone to say that the presence of a risk factor increases the odds of a givenoutcome by a specific factor.
– Logistic regression doesn’t requirelinear relationship between dependent and independent variables. It can handle various types of relationshipsbecause it applies a non-linear log transformation to the predicted odds ratio.(Sachan,2017).
The type of logistic regression1. Binarylogistic regression (Wiley,2011)- used when the dependent variable isdichotomous and the independent variables are either continuous or categorical.- When thedependent variable is not dichotomous and is comprised of more than twocategories, a multinomial logistic regression.2. MultinomialLogistic Regression (Wiley,2011)- The linearregression analysis to conduct when the dependent variable is nominal with morethan two levels. Thus it is an extension of logistic regression, whichanalyses dichotomous (binary) dependents. – Multinomialregression is used to describe data and to explain the relationship between onedependent nominal variable and one or more continuous-level (interval or ratioscale) independent variables.
The logistic regression does not assume a linear relationship betweenthe independent variable and dependent variable and it may handle nonlineareffects. The dependent variable need not be normally distributed. It does notrequire that the independents be interval and unbounded. Logistic regressioncome at a cost, it requires much more data to achieve stable, meaningfulresults. logistic regression come at a cost: it requires much more data toachieve stable, meaningful results. With standard regression, and dependentvariable, typically 20 data points per predictor is considered the lower bound.For logistic regression, at least 50 data points per predictor is necessary toachieve stable results (Wiley,2011) 3) Neural NetworkNeural network is a method of the computing,based on the interaction of multiple connected processing elements.
Ability todeal with incomplete information. When an element of the neural network fails,it can continue without any problem by their parallel nature.(Liu, Yang and Ramsay, 2011) Basic concept of theneural network (Liu, Yang and Ramsay, 2011) 1.Computational Neuroscience- understanding and modelling operations ofsingle neurons or small neuronal circuits, e.g. minicolumns. – Modelling information processing in actualbrain systems, e.
g. auditory tract. – Modelling human perception and cognition. 2.Artificial Neural Networks- Used in Pattern recognition, adaptivecontrol, time series prediction and etc.
– Theareas contributing to Artificial neural networks are Statistical Patternrecognition, Computational Learning Theory, Computational Neuroscience,Dynamical systems theory and Nonlinear optimisation.The type of neuralnetwork (Hinton,2010)1. Feed-Forwardneural network- There is the commonest type of neuralnetwork in practical application. The first layer is the input and the lastlayer is output. – If the is more than one hidden layer, wecall them ‘deep’ neural networks.
They compute a series of transformation thatchange the similarities between cases.2. Recurrentnetworks- These have directed cycles in theirconnection graph. That means you can sometimes get back to where you started byfollowing the arrows.
– They can have complicated dynamic and this canmake them very difficult to train.A neural network can perform tasks that a linear program cannot. A neuralnetwork learns and does not need to be reprogrammed. It can be implemented inany application. It can be implemented without any problem.
Neural networksrequiring less formal statistical training, ability to implicitly detectcomplex nonlinear relationships between dependent and independent