def test(model, X_test, y_test):
loss, accuracy, recall, precision = model.evaluate(X_test, y_test)
print("accuracy: {}%".format(accuracy * 100))
print("recall: {}%".format(recall * 100))
print("precision: {}%".format(precision * 100))
outputs3 = model.predict(X_test)
answers3 = [np.argmax(output) for output in outputs3]
targets3 = [np.argmax(truth) for truth in y_test]
util.analyze(answers3, targets3, [0, 1, 2, 3])
def test(model, X_test, y_test):
loss, accuracy, recall, precision = model.evaluate(X_test, y_test)
print("accuracy: {}%".format(accuracy * 100))
outputs = model.predict(X_test)
answers = [output[0] > 0.5 for output in outputs]
print(answers)
targets = [truth for truth in y_test]
print(targets)
util.analyze(answers, targets, [0, 1])
except (Exception):
print("set_mapping except")
# 主函数
if __name__ == '__main__':
# args = read_args()
# 初始化es环境
init_es(hosts=["localhost:9200"], timeout=5000)
# 创建es类
es = es_tool(hosts=["localhost:9200"], timeout=5000)
# 执行写入操作
batchsize = 50000
i = 0
datalist = []
with open("/data/SDDMCS/LA_orgs/proc.txt") as f:
while i < batchsize:
i += 1
line = f.readline()
if not line:
if len(datalist) != 0:
es.set_data(datalist)
break
datalist.append(analyze(line))
if (i == batchsize):
# print("datalist",datalist)
es.set_data(datalist)
datalist = []
i = 0
y,
num_steps,
data_path,
log_path,
gpu=gpu,
norm=True,
learning_rate=0.1)
#train the model
som.train()
# get the trained map
#net = som.get_weights()
### save/load weights
som.save_weights()
#som.load_weights("./data/test.npy")
# get the location / cluster of the data
data_map, clstr_map = som.map_data()
########################################################
###########
# ANALYZE #
###########
idxs = np.load("./data/idxs.npy")
code_lbl = np.load("./data/code_lbl.npy")
pred_counts, pred_acc, best_pred = analyze(clstr_map, code_lbl, 0.6, 2)