tirthajyoti

# Generate test data
data1 = np.random.normal(loc=5.0, scale=10, size=1000)

# Initialize model
dist1 = distfit(bins=25,alpha=0.02,stats='ks')

# Fit
dist1.fit_transform(data1,verbose=1)

tirthajyoti

object comb {

  def nCr(n: Int, r: Int) = {
    // Factorial
    def fact(i: Int) = {
      var res = 1
      for (e <-i to 2 by -1)
        res*=e
      res
    } // End 'fact' function

tirthajyoti

p1 = drug_user(prob_th=0.5,sensitivity=0.97,specificity=0.95,prevelance=0.005)
print("Probability of the test-taker being a drug user, in the first round of test, is:",round(p1,3))
print()
p2 = drug_user(prob_th=0.5,sensitivity=0.97,specificity=0.95,prevelance=p1)
print("Probability of the test-taker being a drug user, in the second round of test, is:",round(p2,3))
print()
p3 = drug_user(prob_th=0.5,sensitivity=0.97,specificity=0.95,prevelance=p2)
print("Probability of the test-taker being a drug user, in the third round of test, is:",round(p3,3))

tirthajyoti

ps = []
sens = []
for sen in [i*0.001+0.95 for i in range(1,50,2)]:
    sens.append(sen)
    p = drug_user(prob_th=0.5,sensitivity=sen,specificity=0.95,prevelance=0.005,verbose=False)
    ps.append(p)

plt.figure(figsize=(10,5))
plt.title("Probability of user with test sensitivity",fontsize=15)
plt.plot(sens,ps,color='k',marker='o',markersize=8)

tirthajyoti

ps = []
pres = []
for pre in [i*0.001 for i in range(1,51,2)]:
    pres.append(pre*100)
    p = drug_user(prob_th=0.5,sensitivity=0.97,specificity=0.95,prevelance=pre,verbose=False)
    ps.append(p)

plt.figure(figsize=(10,5))
plt.title("Probability of user with prevelance rate",fontsize=15)
plt.plot(pres,ps,color='k',marker='o',markersize=8)

tirthajyoti

def drug_user(
        prob_th=0.5,
        sensitivity=0.99,
        specificity=0.99,
        prevelance=0.01,
        verbose=True):
    """
    Computes the posterior using Bayes' rule
    """
    p_user = prevelance

tirthajyoti

for i,e in enumerate(range(epochs)):
    optimizer.zero_grad()
    output = model.forward(X)
    loss = criterion(output,y)
    loss.backward()
    optimizer.step()
    running_loss.append(loss.item())
    if i!=0 and (i+1)%20==0:
        logits = model.forward(X).detach().numpy().flatten()
        plt.figure(figsize=(15,3))

tirthajyoti

epochs = 10
for i,e in enumerate(range(epochs)):
    optimizer.zero_grad() # Reset the grads
    output = model.forward(X) # Forward pass
    loss = criterion(output.view(output.shape[0]),y) # Calculate loss
    print(f"Epoch - {i+1}, Loss - {round(loss.item(),3)}")  # Print loss
    loss.backward() # Backpropagation
    optimizer.step() # Optimizer one step

tirthajyoti

# Resets the gradients i.e. do not accumulate over passes
optimizer.zero_grad()
# Forward pass
output = model.forward(X)
# Calculate loss
loss = criterion(output,y)
# Backward pass (AutoGrad)
loss.backward()
# One step of the optimizer
optimizer.step()

tirthajyoti

model = Network()
print(model)

Network(
  (hidden1): Linear(in_features=5, out_features=8, bias=True)
  (hidden2): Linear(in_features=8, out_features=4, bias=True)
  (relu): ReLU()
  (output): Linear(in_features=4, out_features=1, bias=True)
  (sigmoid): Sigmoid()
)