smoothts

Smooth data

smoothts is not recommended. Use smoothdata instead.

Syntax

output = smoothts(input)

output = smoothts(input,'e',n)

output = smoothts(input,method)

output = smoothts(___,wsize)

output = smoothts(___,stdev)

output = smoothts(___,n)

Description

output = smoothts(input) smooths the input data using the default Box method with window size, wsize, of 5.

output = smoothts(input,'b',wsize) smooths the input data using the Box (simple, linear) method. wsize specifies the width of the box to be used.

output = smoothts(input,'g',wsize,stdev) smooths the input data using the Gaussian window method.

output = smoothts(input,'e',n) smooths the input data using the Exponential method. n can represent the window size (period length) or alpha. If n > 1, n represents the window size. If 0 < n < 1, n represents alpha, where

$$\alpha =\frac{2}{wsize+1}.$$

If input is a financial time series object, output is a financial time series object identical to input except for contents. If input is a row-oriented matrix, output is a row-oriented matrix of the same length.

output = smoothts(input,method) smooths the input data using a smoothing method.

output = smoothts(___,wsize) smooths the input data using a smoothing method where wsize specifies the width of the box to be used.

output = smoothts(___,stdev) represents the standard deviation of the Gaussian window.

output = smoothts(___,n) smooths the input data using the Exponential method ('e'). n can represent the window size (period length) or alpha. If n > 1, n represents the window size. If 0 < n < 1, n represents alpha, where

$$\alpha =\frac{2}{wsize+1}.$$

Examples

Smooth Data for Financial Times Series Object

1. Create a financial times series (fints) object using dates and data.

   data = [1:6]';
   dates = [today:today+5]';
   tsobj = fints(dates, data)

   Warning: FINTS is not recommended. Use TIMETABLE instead. For more information, see Convert Financial Time Series Objects (fints) to Timetables. 
   > In fints (line 169) 
    
   tsobj = 
    
       desc:  (none)
       freq:  Unknown (0)
   
       {'dates:  (6)'}    {'series1:  (6)'}
       {'01-Sep-2021'}    {[            1]}
       {'02-Sep-2021'}    {[            2]}
       {'03-Sep-2021'}    {[            3]}
       {'04-Sep-2021'}    {[            4]}
       {'05-Sep-2021'}    {[            5]}
       {'06-Sep-2021'}    {[            6]}

2. Use smoothts to smooth the data.

   output = smoothts(tsobj)

    output = 
    
       desc:  Box-smoothed of 
       freq:  Unknown (0)
   
       {'dates:  (6)'}    {'series1:  (6)'}
       {'01-Sep-2021'}    {[       1.2000]}
       {'02-Sep-2021'}    {[       2.0000]}
       {'03-Sep-2021'}    {[       3.0000]}
       {'04-Sep-2021'}    {[            4]}
       {'05-Sep-2021'}    {[       3.6000]}
       {'06-Sep-2021'}    {[            3]}

Input Arguments

input — Input data

fints object | row-oriented matrix

Input data, specified as a fints object or a row-oriented matrix. In a row-oriented matrix, each row represents an individual set of observations.

Data Types: object | double

method — Smoothing method

'b' (default) | character with value of 'b', 'e', or 'g'

Smoothing method, specified as a scalar logical character vector with one of the following values:

• 'b' — Box

• 'e' — Exponential

• 'g' — Gaussian

Data Types: char

wsize — Window size

5 (default) | numeric

Window size, specified as a scalar numeric.

Note

The wsize input argument can only be used when the method is 'b' (Box) or 'g' (Gaussian).

Data Types: double

stdev — Standard deviation of the Gaussian window

0.65 (default) | numeric

Standard deviation of the Gaussian window, specified as a scalar numeric.

Note

The stdev input argument can only be used when the method is 'g' (Gaussian).

Data Types: numeric

n — Window size or exponential factor depending upon value

wsize = 5alpha = 0.3333 (default) | numeric

Window size or exponential factor depending upon value, specified as a scalar numeric with one of the following values:

• n > 1 (window size) or period length

• n < 1 and > 0 (exponential factor: alpha)

• n = 1 (either window size or alpha)

Note

The n input argument can only be used when the method is 'e' (exponential).

Data Types: double

Output Arguments

output — Output

fints object | row-oriented matrix

Output, returned as a fints object or row-oriented matrix.

If input is a financial time series object, output is a financial time series object identical to input except for contents. If input is a row-oriented matrix, output is a row-oriented matrix of the same length.

Version History

Introduced before R2006a