STRLCPY/5G_security_malicious_traffic_detection

Update README.md
Karalis Ilias committed 4 years ago

c886d93f

1 parent c8c087e2

Revision indexing in progress... (symbol navigation in revisions will be accurate after indexed)

Total 1 files

■ ■ ■ ■ ■ ■

README.md

		skipped 79 lines
80	80		* loss = binary_crossentropy
81	81
82	82		DNN results:
83		-	\| precision \| recall \| f1-score \| support \|
	83	+
	84	+	precision recall f1-score support
	85	+
84	86		0 1.00 1.00 1.00 12833
85	87		1 1.00 1.00 1.00 9711
86	88
		skipped 1 lines
88	90		macro avg 1.00 1.00 1.00 22544
89	91		weighted avg 1.00 1.00 1.00 22544
90	92
91		-	Best score: 0.9999680776451911
92		-	Estimator time: 645.5822646617889
93		-	Prediction time: 0.7374951839447021
94		-	Total time for fit and predict: 646.3197634220123 seconds
	93	+	* Best score: 0.9999680776451911
	94	+	* Estimator time: 645.5822646617889
	95	+	* Prediction time: 0.7374951839447021
	96	+	* Total time for fit and predict: 646.3197634220123 seconds
95	97
96	98
97	99		![](plots/cm_nn.png)
		skipped 1 lines
99	101
100	102		NAD NN Network description and results:
101	103		LSTM : Long short term memory recurrent neural network.
102		-	RandomSearchCV was used for training with cross validation and best parameters
103		-	hidden_layers =4
	104	+	RandomSearchCV was used for training with cross validation and best parameters:
	105	+	hidden_layers =4,
104	106		learning_rate = 0.0002267
105	107
	108	+	Training data:
	109	+
	110	+	* Epochs = 250
	111	+	* verbose =2
	112	+	* batch size = 2000
	113	+	* scoring = accuracy
	114	+	* optimizer = Adam
	115	+	* loss = binary_crossentropy
	116	+
	117	+
106	118		LSTM results:
	119	+
107	120		precision recall f1-score support
108	121
109	122		0 1.00 1.00 1.00 12833
		skipped 3 lines
113	126		macro avg 1.00 1.00 1.00 22544
114	127		weighted avg 1.00 1.00 1.00 22544
115	128
116		-	Best score: 0.9998968032832433
117		-	Estimator time: 956.7115938663483
118		-	Prediction time: 1.4246132373809814
119		-	Total time for fit and predict: 958.1362102031708 seconds
120		-	Confusion matrix:
121		-	[[12825 8]
122		-	[ 51 9660]]
	129	+	* Best score: 0.9998968032832433
	130	+	* Estimator time: 956.7115938663483
	131	+	* Prediction time: 1.4246132373809814
	132	+	* Total time for fit and predict: 958.1362102031708 seconds
	133	+
	134	+
123	135		![](plots/cm_lstm.png)

Update README.md