Research method
Autocorrelation test
Autocorrelation test is a canonical archive-taxonomy method; this registry entry defines its inputs, output and use boundary.
- Output
- Conditional forecast or research filter
- Input
- Ordered price, volume or breadth observations
- Horizon
- A defined sampling interval and lookback
- Best use
- Compare an explicit quantitative baseline with an out-of-sample result.
- Input: Ordered price, volume or breadth observations.
- Output: forecast.
- Use boundary: Compare an explicit quantitative baseline with an out-of-sample result.
The graph2
How the method is built, and where it has been read
Depends on
Often reviewed together
Runs test5 readings
Often reviewed together
Chi-square test7 readings
Read in the archive23 readings
1982Construct ARIMA forecasts from lag diagnostics1986Evaluating daily price and volume serial independence windows1986Weekly resamples as a diagnostic filter for statistical windows1987Evaluating weekly price cycles with autocorrelation cutoffs1987Evaluating money-supply serial dependence before a forecast1988Serial correlation is not an independence test1988Evaluating stationarity, randomness, and dependence in an index series1988Event-triggered chi-square tests for jittered serial links1988Reverse the rule before you trust the system1989Statistical windows for indicator time parameters1990Sign-cell correlation, chi-square checks, and lag shifts1990Two-gate out-of-sample tests for trading rules1990Chi-square evaluation of lagged breadth correlations1990Constructing a short-horizon ARIMA from differenced wheat closes1998Test serial dependence before you optimize a mechanical system2003How a one-rank rotation is constructed for an offset autocorrelation check2004A two-stage Durbin-Watson residual screen2004Constructing first-order residual autocorrelation tests2005Evaluating persistence with runs and autocorrelation2011Constructing trend labels from increment correlation atoms2012Overnight versus daytime return evaluation2012Constructing a multi-horizon folded-correlation reversal filter2016Tuning indicators to measured market cycles
This registry entry describes a research method. It is not investment advice and does not make a performance claim.