Nadaraya-Watson: Rational Quadratic Kernel (Non-Repainting)

• 08/31/2023 at 7:25 AM #220102
  Knox

  Hi All,

  I was wondering if it is possible to replicate an indicator from TradingView.

  It is called Nadaraya-Watson: Rational Quadratic Kernel (Non-Repainting). On the charts, it kind of looks like a regular Moving Average. But you can set the lines based on another timefrime

  You can find more info and the code on TV

  Thank you

  

  Knox

  Participant

  Topics: 1

  Replies: 7

  Been thanked: 0 times
  08/31/2023 at 7:44 AM #220103
  JC_Bywan

  Hi,

  translated here by Nicolas last year:

  https://www.prorealcode.com/topic/conversion-from-trading-view-nadaraya-watson-envelope/#post-196311

   

  

  JC_Bywan

  Moderator

  Topics: 20

  Replies: 2651

  Been thanked: 915 times
  08/31/2023 at 8:33 AM #220105
  Knox

  Hi,

  thanks. but that looks like the Envelope and not the Rational Quadratic Kernel (Non-Repainting), no?

  

  Knox

  Participant

  Topics: 1

  Replies: 7

  Been thanked: 0 times
  08/31/2023 at 8:56 AM #220107
  JC_Bywan

  Sorry, I recognised the name and gave the forum link assuming you were after the mid-line (the y2 variable of the return line), if what you’re after is not the same as this y2 then I don’t remember other Nadaraya-Watson code already made in the forum, Nicolas will assess whether it can be converted (and if yes will add it to the conversions list).

  

  JC_Bywan

  Moderator

  Topics: 20

  Replies: 2651

  Been thanked: 915 times
  08/31/2023 at 9:25 AM #220108
  Nicolas

  It is the same as the y2 line from the above code, apart that in that version the line takes color following another 2 period lag line crossing.

  Added to the conversion list in order to get the same as this one.

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  08/31/2023 at 7:03 PM #220150
  Knox

  Ok, so it is not possible to create the same indicator as I posted initially?

  

  Knox

  Participant

  Topics: 1

  Replies: 7

  Been thanked: 0 times
  09/01/2023 at 10:38 AM #220172
  Nicolas

  Yes it is possible, please be patient 😉

  

  Nicolas

  Keymaster

  Topics: 81

  Replies: 20863

  Been thanked: 4773 times
  09/01/2023 at 1:13 PM #220180
  LaMaille

  Done!

  NadarayaWatsRationalQuadKernel

  // © jdehorty // @version=5 // A non-repainting implementation of Nadaraya-Watson Regression using a Rational Quadratic Kernel. // Settings once bullishR = 0 once bullishG = 250 once bullishB = 154 once bearishR = 255 once bearishG = 69 once bearishB = 0 src = customclose // hh = Lookback Window, minval=3.0, tooltip='The number of bars used for the estimation. This is a sliding value that represents the most recent historical bars. Recommended range: 3-50 r = rr*0.25 //rr = Relative Weighting, step=0.25, Relative weighting of time frames. As this value approaches zero, the longer time frames will exert more influence on the estimation. As this value approaches infinity, the behavior of the Rational Quadratic Kernel will become identical to the Gaussian kernel. Recommended range: 0.25-25 => RR from 1 to 100 //xx = 25, Start Regression at Bar, Bar index on which to start regression. The first bars of a chart are often highly volatile, and omission of these initial bars often leads to a better overall fit. Recommended range: 5-25 //smoothColors = false, Smooth Colors, tooltip=Uses a crossover based mechanism to determine colors. This often results in less color transitions overall //lag = 2, Lag for crossover detection. Lower values result in earlier crossovers. Recommended range: 1-2 //size = array.size(array.from(src)) // size of the data series size = Min(Barindex,Maxlength) // Further Reading: // The Kernel Cookbook: Advice on Covariance functions. David Duvenaud. Published June 2014. // Estimation of the bandwidth parameter in Nadaraya-Watson kernel non-parametric regression based on universal threshold level. Ali T, Heyam Abd Al-Majeed Hayawi, Botani I. Published February 26, 2021. // Estimations currentWeight = 0 cumulativeWeight = 0 for i = 0 to Size-xx-1 do y = src[i] w = POW(1+(POW(i,2)/(POW(hh,2)*2*r)),-r) currentWeight = currentWeight + y*w cumulativeWeight = cumulativeWeight + w Next yhat1 = currentWeight / cumulativeWeight currentWeight = 0 cumulativeWeight = 0 for i = 0 to Size-xx-1 do y = src[i] w = POW(1+(POW(i,2)/(POW(hh-lag,2)*2*r)),-r) currentWeight = currentWeight + y*w cumulativeWeight = cumulativeWeight + w Next yhat2 = currentWeight / cumulativeWeight // Rates of Change wasBearish = yhat1[2] > yhat1[1] wasBullish = yhat1[2] < yhat1[1] isBearish = yhat1[1] > yhat1 isBullish = yhat1[1] < yhat1 isBearishChange = isBearish and wasBullish isBullishChange = isBullish and wasBearish // Crossovers isBullishCross = yhat2 crosses over yhat1 isBearishCross = yhat2 crosses under yhat1 isBullishSmooth = yhat2 > yhat1 isBearishSmooth = yhat2 < yhat1 // Colors If isBullishSmooth then colorByCrossR = bullishR colorByCrossG = bullishG colorByCrossB = bullishB Else colorByCrossR = bearishR colorByCrossG = bearishG colorByCrossB = bearishB Endif If isBullish then colorByRateR = bullishR colorByRateG = bullishG colorByRateB = bullishB Else colorByRateR = bearishR colorByRateG = bearishG colorByRateB = bearishB Endif If smoothColors then R = colorByCrossR G = colorByCrossG B = colorByCrossB Else R = colorByRateR G = colorByRateG B = colorByRateB Endif // Alert Variables If smoothColors then alertBullish = isBearishCross Else alertBullish = isBearishChange Endif If smoothColors then alertBearish = isBullishCross Else alertBearish = isBullishChange Endif // Alerts for Color Changes //alertcondition(condition=alertBullish, title='Bearish Color Change', message='Nadaraya-Watson: {{ticker}} ({{interval}}) turned Bearish ▼') //alertcondition(condition=alertBearish, title='Bullish Color Change', message='Nadaraya-Watson: {{ticker}} ({{interval}}) turned Bullish ▲') // Non-Displayed Plot Outputs (i.e., for use in other indicators) //plot(alertBearish ? -1 : alertBullish ? 1 : 0, "Alert Stream", display=display.none) Return yhat1 as "Rational Quadratic Kernel Estimate" style(line,4) coloured(R,G,B,255)

  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121

  // © jdehorty // @version=5   // A non-repainting implementation of Nadaraya-Watson Regression using a Rational Quadratic Kernel.   // Settings once bullishR = 0 once bullishG = 250 once bullishB = 154 once bearishR = 255 once bearishG = 69 once bearishB = 0   src = customclose   // hh = Lookback Window, minval=3.0, tooltip='The number of bars used for the estimation. This is a sliding value that represents the most recent historical bars. Recommended range: 3-50   r = rr*0.25 //rr = Relative Weighting, step=0.25, Relative weighting of time frames. As this value approaches zero, the longer time frames will exert more influence on the estimation. As this value approaches infinity, the behavior of the Rational Quadratic Kernel will become identical to the Gaussian kernel. Recommended range: 0.25-25 => RR from 1 to 100   //xx = 25, Start Regression at Bar, Bar index on which to start regression. The first bars of a chart are often highly volatile, and omission of these initial bars often leads to a better overall fit. Recommended range: 5-25   //smoothColors = false, Smooth Colors, tooltip=Uses a crossover based mechanism to determine colors. This often results in less color transitions overall   //lag = 2, Lag for crossover detection. Lower values result in earlier crossovers. Recommended range: 1-2   //size = array.size(array.from(src)) // size of the data series size = Min(Barindex,Maxlength)   // Further Reading: // The Kernel Cookbook: Advice on Covariance functions. David Duvenaud. Published June 2014. // Estimation of the bandwidth parameter in Nadaraya-Watson kernel non-parametric regression based on universal threshold level. Ali T, Heyam Abd Al-Majeed Hayawi, Botani I. Published February 26, 2021.   // Estimations currentWeight = 0 cumulativeWeight = 0 for i = 0 to Size-xx-1 do y = src[i] w = POW(1+(POW(i,2)/(POW(hh,2)*2*r)),-r) currentWeight = currentWeight + y*w cumulativeWeight = cumulativeWeight + w Next yhat1 = currentWeight / cumulativeWeight   currentWeight = 0 cumulativeWeight = 0 for i = 0 to Size-xx-1 do y = src[i] w = POW(1+(POW(i,2)/(POW(hh-lag,2)*2*r)),-r) currentWeight = currentWeight + y*w cumulativeWeight = cumulativeWeight + w Next yhat2 = currentWeight / cumulativeWeight   // Rates of Change wasBearish = yhat1[2] > yhat1[1] wasBullish = yhat1[2] < yhat1[1] isBearish = yhat1[1] > yhat1 isBullish = yhat1[1] < yhat1 isBearishChange = isBearish and wasBullish isBullishChange = isBullish and wasBearish   // Crossovers isBullishCross = yhat2 crosses over yhat1 isBearishCross = yhat2 crosses under yhat1 isBullishSmooth = yhat2 > yhat1 isBearishSmooth = yhat2 < yhat1   // Colors   If isBullishSmooth then colorByCrossR = bullishR colorByCrossG = bullishG colorByCrossB = bullishB Else colorByCrossR = bearishR colorByCrossG = bearishG colorByCrossB = bearishB Endif   If isBullish then colorByRateR = bullishR colorByRateG = bullishG colorByRateB = bullishB Else colorByRateR = bearishR colorByRateG = bearishG colorByRateB = bearishB Endif   If smoothColors then R = colorByCrossR G = colorByCrossG B = colorByCrossB Else R = colorByRateR G = colorByRateG B = colorByRateB Endif   // Alert Variables If smoothColors then alertBullish = isBearishCross Else alertBullish = isBearishChange Endif   If smoothColors then alertBearish = isBullishCross Else alertBearish = isBullishChange Endif   // Alerts for Color Changes //alertcondition(condition=alertBullish, title='Bearish Color Change', message='Nadaraya-Watson: {{ticker}} ({{interval}}) turned Bearish ▼') //alertcondition(condition=alertBearish, title='Bullish Color Change', message='Nadaraya-Watson: {{ticker}} ({{interval}}) turned Bullish ▲')   // Non-Displayed Plot Outputs (i.e., for use in other indicators) //plot(alertBearish ? -1 : alertBullish ? 1 : 0, "Alert Stream", display=display.none)   Return yhat1 as "Rational Quadratic Kernel Estimate" style(line,4) coloured(R,G,B,255)

  2 users thanked author for this post.

  JS, Distincttacos

  

  LaMaille

  Participant

  Topics: 1

  Replies: 18

  Been thanked: 12 times
  09/01/2023 at 3:02 PM #220187
  Knox

  Haha ok will do. Thank you!

  

  Knox

  Participant

  Topics: 1

  Replies: 7

  Been thanked: 0 times
  09/01/2023 at 3:02 PM #220188
  Knox

  Oh wow, many thanks. Will try it over the weekend.

  

  Knox

  Participant

  Topics: 1

  Replies: 7

  Been thanked: 0 times
  09/01/2023 at 3:42 PM #220189
  Knox

  I get this error message. On which rows am I suppose to add values?

  sorry, really new to coding

   

  thank you

  

  Knox

  Participant

  Topics: 1

  Replies: 7

  Been thanked: 0 times
  09/01/2023 at 4:05 PM #220192
  JS

  Hi @Harem

  Download the itf -file and import the itf-file under indicators…

  Now you have the complete code including the parameters…

  (So, don’t copy the code from the forum but download the itf-file)

  

  JS

  Participant

  Topics: 26

  Replies: 1716

  Been thanked: 779 times
  09/01/2023 at 4:25 PM #220193
  Knox

  much appreciated. Thank you

  

  Knox

  Participant

  Topics: 1

  Replies: 7

  Been thanked: 0 times
  09/02/2023 at 5:25 PM #220225
  Knox

  Hi,

  sorry, one more question. Do I have to have the maxlength parameter? or is there a default nr for that?

  

  Knox

  Participant

  Topics: 1

  Replies: 7

  Been thanked: 0 times
  09/03/2023 at 1:25 PM #220249
  LucasBest

  I added Maxlength because otherwise the indicator is very slow.
  I just translate the original code (without improving it to be faster, which is possible) and it seems it uses all barindex for calculation (which is not relevant too because it does not change the result after 200 or 300 bars).
  That is why i added Maxlength (too lazzy for improving the code)
  But you can change 400 to 2000 or more if you want, you will notice no change in the result exept more time of calculation

  ps : Lamaille is me

  

  LucasBest

  Participant

  Topics: 16

  Replies: 208

  Been thanked: 72 times
• Author
  Posts