Basically, a Cluster Indicator should be used in a trading strategy as a buy rule, buying when the current bar is "close" enough to the cluster center of recent "buy" bars.  Likewise, you should sell when the current bar is "close" enough to the "sell" cluster center.

You put one Cluster Indicator in as a long entry condition using the relational indicator (A<B), where A is the Cluster Indicator, and B is a numeric threshold. Then you put another one in as a long exit (or short entry) condition, with A<B, again where A is the Cluster Indicator, and B is a numeric threshold. Then the genetic algorithm optimizer in the NeuroShell Trader Pro or DayTrader Pro finds the cluster centers and threshold values that optimize the profit.

Note 1: The numeric threshold B should vary from 0 to 1 during optimization.

Note 2: The relation indicator A<B is used both with a buy rule and a sell rule. We do NOT use A>B.

Note 3: If you have a Cluster Indicator for a long entry, a long exit, a short entry, and a short exit, they are all optimized independently. Therefore, there are really four cluster centers being determined, which is appropriate.

Note 4:  We believe that using Cluster Indicators with fewer inputs will more likely result in models that are more robust when applied to future data.

The Cluster Indicators are not predicting any value into the future. They are simply giving "intelligent" signals based on what they "know" about the past. The indicator is only optimized to give buy/sell signals.

Each of the Cluster Indicators will have inputs, which are usually some other regular indicators. You might, for example, like to use the RSI, CCI, and DMI indicators as inputs to each of the Cluster Indicators. A Cluster Indicator is returning the distance from the cluster center to the current bar. The current bar is defined by the current values of the input indicators. The cluster center is defined by the parameters C1, C2, etc. of the Cluster Indicator. The optimizer finds the values of C1, C2, etc.

`