# Train: Train In SmartSifter: Online Unsupervised Outlier Detection Using Finite Mixtures with Discounting Learning Algorithms

## Description

trains the parameters

## Usage

 `1` ```Train(y, param = TRUE, const, initial) ```

## Arguments

 `y` A matrix, the training set. `param` A logical scalar, if TRUE, the model is in parametric version, otherwise, a non-parametric one. `const` A numeric vector, specifies the value of all global variables, if param = T, then const = c(N,n,d,rh,r,K,alpha); if param=FALSE, then const = c(N,n,d,rh,r,K,sigma_sqare). `initial` A numeric vector, specifies the initial value of parameters over the continuous domain, if param = T, initial = c(pi_1,mean_1,cov_1, ..., pi_K, mean_K,cov_K), if param = F, initial = c(q1,q2, ..., qK).

## Value

List of all parameters.

## Examples

 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19``` ```##parametric model test initial=matrix(c(0.5,0,0,1,0,0,1,0.5,1,1,1,0,0,1),nrow=1) const = c(0,1,2,0.1,0.1,2,2) param=TRUE y=matrix(c(1,3,1,0,1,1),nrow=2) smart = Train(y,param,const,initial)\$smart hellingerScore = Train(y,param,const,initial)\$HellingerScore logLoss = Train(y,param,const,initial)\$LogLoss const[1] = Train(y,param,const,initial)\$N ##non-parametric model test param=FALSE const = c(0,1,2,0.1,0.1,2,1) initial = matrix(c(0,0,1,1),nrow=1) y=matrix(c(1,3,1,0,1,1),nrow=2) smart = Train(y,param,const,initial)\$smart hellingerScore = Train(y,param,const,initial)\$HellingerScore logLoss = Train(y,param,const,initial)\$LogLoss const[1] = Train(y,param,const,initial)\$N ```

### Example output

```
```

SmartSifter documentation built on May 29, 2017, 9:12 p.m.