Guys
The stock market is rather tricky and driven by human psychology of fear, anxiety and greed. What triggers a human being at a particular time cannot be predicted.
So my model does not give stop losses and others - it just looks at the tendency of the stock - if the current state of equilibrium persists. From there - it gives you three states where it could go - an aggressive state , if market moves aggressively, a conservative state and stable market state....
You have to understand which state the market is in....
For example, my model is telling not to sell INDIGO right now...as stocks are going up and up. So what happened in with the Suzlon news coming out - the market was falling. In this case, to raise the markets - some stocks are inflated and some are deflated. It did not predict the 1300 levels - but did predict it going down....What happened was an SEC violation - but that's a different story. But you still could have made money in intra-day or could have held it with e-margins that I play on...
KNN means K-nearest Neighbours....so the algorithm finds the nearest distances to stock indices around the world. For example, an IOC stock will be closer to international oil prices, a steal stock would be closer to steal prices.
And when market was falling - did I not say that it would reach 11691...look where is it now when most people were expecting it to come down.
What kind of fluctuations....distance of stock to indices u mean....which indices....
As for indigo u said u are predicting reaching 1510-20 levels....did it also predict fall to 1300 odd levels...
We all here are trying our best to learn to trade markets.....based on our individual potential, capital employed and risk taking capacity....
If you bring an algorithm that can provide an edge to market participants, most welcome but when you put statements like following indices will go to this level, following stock will go to this level without an time horizon or stop loss, what does one call it...
U are right I don't understand artificial intelligence algorithm but am open to learn new techniques...