Fire (TsFile 1.2.1 API)

java.lang.Object
- org.apache.iotdb.tsfile.encoding.fire.Fire<T>

Direct Known Subclasses:

IntFire, LongFire
```
public abstract class Fire<T extends java.lang.Comparable<T>>
extends java.lang.Object
```
This class is used for Fast Integer REgression in Sprintz encoding method.Users can create a Fire object and use it to train and predict the time-ordered integer-like data.

Field Summary

Fields
Modifier and Type Field Description

protected int accumulator

protected int bitWidth

protected T delta

protected int learnShift

Constructor Summary

Constructors
Constructor Description

Fire(int learning_rate)

Method Summary

All Methods Instance Methods Abstract Methods
Modifier and Type	Method	Description
`abstract T`	`predict(T value)`	predict the incoming integer using the last value
`abstract void`	`train(T pre, T val, T err)`	train the learning machine with the last prediction

Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

learnShift
```
protected int learnShift
```

bitWidth
```
protected int bitWidth
```

accumulator
```
protected int accumulator
```

delta

protected T extends java.lang.Comparable<T> delta

Constructor Detail
- Fire
```
public Fire(int learning_rate)
```

Method Detail
- predict
```
public abstract T predict(T value)
```
  predict the incoming integer using the last value
  
  Parameters:
  
  value - the last value
  
  Returns:
  
  the predicted value
- train
```
public abstract void train(T pre,
                           T val,
                           T err)
```
  train the learning machine with the last prediction
  
  Parameters:
  
  pre - last value to be predicted
  
  val - current value to be predicted
  
  err - the predictive error of current value