Five examples of how to use Cluster Indicators are called Fuzzy Example 1.cht, Fuzzy Example 2, and so on. The charts for the examples are installed in the "Fuzzy Examples" subfolder of the NeuroShell Trader folder.
Fuzzy Example 1
In Fuzzy Example 1, we used BCC (Boise Cascade Co.) daily stock prices. We employed the Fuzzy2 indicator to search for the "Close rises, then drops" pattern with 8 bars in each segment. This is precisely the same chart we have used to draw Fig.1 in the "Parameters and Recommended Settings: Fuzzy Indicators" topic of this help file. Please refer to this topic for a more detailed analysis of this example chart.
Fuzzy Example 2
In Fuzzy Example 2, we used IBM daily stock prices. We built a simple trading strategy in which the same Fuzzy3 indicator was used in a long entry condition (rule), and another for short entry condition. There were no exit conditions; we reversed from a long to a short and vice versa. The actual conditions for both the long entry and the short entry were the same: the relational A>B indicator, where A was the Fuzzy3 indicator and B was allowed to vary from 0 to 1. It is worth pointing out here that although the same A>B was selected for both the buy and sell rules, the trading strategy actually uses two copies of Fuzzy3 which it evolves differently. One evolves to predict long entry predictions, and another evolves to predict short entry predictions.
The Close price was used as the input time series. We used the optimizer in the NeuroShell Trader Professional to find the pattern rules (Code1, Code2, and Code3), as well as the segment size, the Max Change parameter, and the B parameter. We optimized over two years, and then backtested over the next six months. The goal was to maximize the return on account with starting capital equivalent to 100 shares.
The Short Entry Rule and Long Entry Rule on the chart show the optimized Fuzzy3 indicator outputs. It is interesting to note that the Long entry pattern found is "Close drops, then rises, then drops again" with 16 bars in each segment. The Short entry pattern is "Close sharply rises, then sharply drops, then sharply rises again" with 22 bars in each segment.
The results were a substantial improvement over buy and hold for both in-sample (optimization period) and out-of-sample (backtest period).
Fuzzy Example 3
Fuzzy Example 3 is virtually the same as Fuzzy Example 2 except that we used the Fuzzy5 instead of the Fuzzy3 indicator.
Again, the results were a substantial improvement over buy and hold for both in-sample (optimization period) and out-of-sample (backtest period).
Fuzzy Example 4
Fuzzy Example 4 is very much like the second and third examples with the following major exceptions:
1. We used four more stock prices (BCC, CHV, DE, EK) in addition to IBM.
2. We used FuzzyGA3 indicator instead of Fuzzy3 or Fuzzy5 for both the long entry and short entry rules.
As you might notice, we prefer stocks which tend to oscillate without the dominance of bull or bear markets.
Again, the results were a substantial improvement over buy and hold for both in-sample (optimization period) and out-of-sample (backtest period).
Fuzzy Example 5
Fuzzy Example 5 is an example of using Fuzzy Rules to create more complex fuzzy rules out of elementary ones. We used the same BCC (Boise Cascade Co.) price time series as in the Fuzzy Example 1. Our goal here is to find combined occurrences of EITHER two patterns: one is "Close rises, then drops", another one is "High remains steady, then drops, then remains steady again, then rises". We do this by using the FuzzyOR2 indicator, which takes as two parameters the elementary Fuzzy2 and Fuzzy4 indicators.
Each Fuzzy2 and Fuzzy4 takes its own price time series and own set of fuzzy logic parameters. The Fuzzy2 indicator looks for the "Close rises, then drops", the Fuzzy4 indicator looks for the "High remains steady, then drops, then remains steady again, then rises" pattern. The final FuzzyOR2 indicator rule output is the logical OR combination of the Fuzzy2 and Fuzzy4 indicators outputs.
|