The Adaptive-Clustering and Error-Correction Method for Forecasting Cyanobacteria Blooms in Lakes and Reservoirs

Table 3

Comparison of results.

Prediction method	Evaluation index
	Average absolute error	Average relative error
Phase space reconstruction	0.4440	0.3365
Traditional-clustering linear regression	0.1887	0.1368
Adaptive-clustering linear regression (Ht index)	0.1818	0.1443
Adaptive-clustering linear regression (KL index)	0.1380	0.1204
Adaptive-clustering linear regression (CH index)	0.1456	0.1231
Adaptive-clustering linear regression (clustering comprehensive quality)	0.1553	0.1171
Adaptive-clustering fuzzy linear regression (clustering comprehensive quality)	0.1276	0.0994