def update(val):
gamma = sgamma.val
alpha = salpha.val
beta = sbeta.val
G = sG.val
n = min(sn.val, G)
nsteps = int(snsteps.val)
lambdaval = int(slambda.val)
deltaval = int(sdelta.val)
# re-do the plot
ax = subplot(111)
data = loadGroupData.load_data(data_N)
t_data = range(len(data[0]))
# clear old plot
cla()
yvals = myfun(gamma, alpha, beta, G, n, nsteps, lambdaval, deltaval)
#t = range(len(yvals))
#t = [ float(item)/nsteps for item in t]
t = range(len(yvals[0]))
# assert len(yvals)==len(t), "unequal lengths for plotting"
#print '----------------------------'
#print yvals
plot(t_data, data[3], 'r+')
plot(t_data, data[1], 'b+')
plot(t, yvals[0], lw=2, color='blue')
plot(t, yvals[1], lw=2, color='black')
plot(t, yvals[2], lw=2, color='red')
ptitle= "Wechat Group g=%2.2f a=%2.2f b=%2.2f init=%2.2f , nsteps=%3d" % \
(gm0, a0, b0, n0/G0, nsteps0)
title(ptitle)
xlabel("time")
ylabel("value")
axis([0, nsteps, 0, G + 50])
draw()
csvfile = file('../../rawresult/simpleSIR.csv', 'rb')
reader = csv.reader(csvfile)
params = list()
comermse = list()
atrmse = list()
gormse = list()
for line in reader:
params.append(line)
test = list()
total = 135621
for i in range(total):
#print i
if i % 1000 == 0:
print i
rawdata = loadGroupData.load_data(i)
at = list()
for j in range(length):
at.append(float(rawdata[1][j]) - float(rawdata[3][j]))
come = numpy.array(rawdata[1][:length])
go = numpy.array(rawdata[3][:length])
result = list()
#res = getCurve(params[i], 0)
t1, t2 = getCurve(params[i], 0, len(come))
temp = rmse(t1, come)
s1 += temp
result.append(temp)
come = [k/float(params[i][2]) for k in come]
temp = rmse(t2, come)
result.append(temp)
import loadGroupData
import optGroupData
import disGroupData
from lmfit import report_fit
traindays = 20
begin = 0
end = 135621
mode = 0 #[0, 1, 2, 3, 4, 5] for [multi, single, SIRd, SIR, SI, SpikeM]
name = ['multi', 'single', 'SIRd', 'SIR', 'SI']
fw = open('../../' + name[mode] + '_res/params.csv', 'w')
for i in range(begin, end):
rawdata = loadGroupData.load_data(i)
result = optGroupData.optimize(rawdata, mode)
report_fit(result.params)
fw.write(str(i) + ',')
fw.write(str(result.params))
fw.write('\n')
path = '../../' + name[mode] + '/'
#disGroupData.display_bi(rawdata,result,result_test,path+str(i)+'.png')
p = disGroupData.display(rawdata, result, path + str(i) + '.png', mode)
if peaks.has_key(p):
peaks[p] += 1
else:
peaks[p] = 1
print 'No.' + str(i) + ' items finished.(' + name[i] + ')'
fw.close()
print peaks
b0 = 0.05 # beta default
G0 = 300 # matches the Glass-Mackey for white blood cells
n0 = 10 # matches GM
nsteps0 = 50
lambda0 = 0
delta0 = 1
s = myfun(gm0, a0, b0, G0, n0, nsteps0, lambda0, delta0)
# create the horizontal axis
t = range(len(s[0]))
print "duration = ", len(s) / nsteps0
#t = [ float(item)/nsteps0 for item in t]
# l, = plot(t,s, lw=2, color='red')
# l = plot(t,s, lw=2, color='red')
data_N = 8
data = loadGroupData.load_data(data_N)
t_data = range(len(data[0]))
plot(t_data, data[0], 'b+')
plot(t_data, data[1], 'r+')
plot(t, s[0], lw=2, color='blue')
plot(t, s[1], lw=2, color='black')
plot(t, s[2], lw=2, color='red')
xlabel("time")
ylabel("value")
ptitle= "Wechat Group g=%2.2f a=%2.2f b=%2.2f init=%2.2f , nsteps=%3d" % \
(gm0, a0, b0, n0/G0, nsteps0)
title(ptitle)
axis([0, nsteps0, 0, G0 + 50])
import loadGroupData
import optGroupData
import disGroupData
from lmfit import report_fit
traindays = 20 #how many days used to train(42 for all)
begin = 0 #start with which group
end = 135622 #end with which group
mode = 0 #[0, 1, 2, 3, 4] for [multi, single, SIRd, SIR, SI]
name = ['multi', 'single', 'SIRd', 'SIR', 'SI']
fw=open('../../' + name[mode] + '_res/params.csv','w')
for i in range(begin, end):
rawdata=loadGroupData.load_data(i, traindays)
result=optGroupData.optimize(rawdata, mode)
report_fit(result.params)
fw.write(str(result.params))
fw.write('\n')
path='../../'+name[mode]+'/'
#disGroupData.display_bi(rawdata,result,result_test,path+str(i)+'.png')
print 'No.' + str(i) + ' items finished.('+name[i]+')'
fw.close()
print peaks
def getItem(order, mode):
rawdata = loadGroupData.load_data(i)
result = optGroupData.optimize(rawdata, mode)
return str(result.params)
