「ぎょーむ日誌」目次に戻る | KuboWeb top に戻る | twilog | atom

ぎょーむ日誌 2002-08-28

苦情・お叱りは, たいへんお手数かけて恐縮ですが, 久保 (kubo@ees.hokudai.ac.jp) までお知らせください.

2002 年 08 月 28 日 (水)

5 N 個パラメーターモデル {A} {B} {C} {D} {E} → あてはまりの良さ I5N
最良の 4 N 個パラメーターモデル {A+C} {B} {D} {E} → あてはまりの良さ I4N
最良の 3 N 個パラメーターモデル {A+C} {B} {D+E} → あてはまりの良さ I3N
……
	-----------------------------------------------------------------
	# model_group = F1,F2,S1,S2,S3
	number_of_parameter = 10
	log_likelihood      = -9.0332e+01
	AIC                 = 2.0066e+02
	[F1]
	  log_likelihood = -1.5421e+01, parameters =  -7.28e+00  3.98e-01
	[F2]
	  log_likelihood = -1.9849e+01, parameters =  -7.19e+00  3.47e-01
	[S1]
	  log_likelihood = -1.7984e+01, parameters =  -6.66e+00  3.26e-01
	[S2]
	  log_likelihood = -2.0512e+01, parameters =  -5.88e+00  2.73e-01
	[S3]
	  log_likelihood = -1.6566e+01, parameters =  -1.13e+01  5.20e-01
	-----------------------------------------------------------------
	…… (中略) ……
	-----------------------------------------------------------------
	# model_group = F1,F2+S1,S2+S3
	number_of_parameter = 6
	log_likelihood      = -1.0459e+02
	AIC                 = 2.2117e+02
	[F1]
	  log_likelihood = -1.5421e+01, parameters =  -7.28e+00  3.98e-01
	[F2+S1]
	  log_likelihood = -3.8082e+01, parameters =  -6.92e+00  3.36e-01
	[S2+S3]
	  log_likelihood = -5.1083e+01, parameters =  -7.68e+00  3.56e-01
	-----------------------------------------------------------------

	#the best model set is
	-----------------------------------------------------------------
	# model_group = F1,F2+S1,S2,S3
	number_of_parameter = 8
	log_likelihood      = -9.0581e+01
	AIC                 = 1.9716e+02
	[F1]
	  log_likelihood = -1.5421e+01, parameters =  -7.28e+00  3.98e-01
	[F2+S1]
	  log_likelihood = -3.8082e+01, parameters =  -6.92e+00  3.36e-01
	[S2]
	  log_likelihood = -2.0512e+01, parameters =  -5.88e+00  2.73e-01
	[S3]
	  log_likelihood = -1.6566e+01, parameters =  -1.13e+01  5.20e-01
	-----------------------------------------------------------------

KuboLog | KuboWeb