See also questions close to this topic

• ordinary least squares coefficients calculation using np.einsum

  Is there a way to calculate ordinary least squares coefficients β in one np.einsum call given a vector of dependent variable Y and a matrix of predictors X?
• System of coupled ODE - heat exchanger problem

  I'm solving a heat exchanger problem and have to find the final temperature. It's a system of ode's I've written the following code: It asks for me to define T and t, but I they're functions of x that I should find

  import numpy as np
  import matplotlib.pyplot as plt
  from scipy.integrate import odeint
  
  #propriedades trietileno glicol na entrada
  T1 = 363 #K temperatura de entrada
  Wt = 3.6 #kg/s vazão
  
  #propriedades água na entrada
  t1 = 298 #K 
  Ww = 1.0
  Ww = [0.1, 0.5, 1] #kg/s vazões
  
  Di = 0.07792 #m diâmetro interno
  L = 3.048 #m comprimento do tubo
  
  U = 283.72 #W/m²K
  pi= 3.1415
  
  cpt = 1901.09#, 2.7683*T + 896.2] #J/kg*K calor específico
  cpw = 4018.04#, -0.00003*(t^3) + 0.0403*(t^2) - 16.277*t + 6083.7] #J/kg*K calor específico
  cp = [cpt, cpw]
  
  def funct(y, x):   
      y = T, t
      dTdx = (U * Di * pi / (Wt * cpt)) * (T - t)
      dtdx = (U * Di * pi / (Ww * cpw)) * (T - t)
      x = np.linspace(0, L, 100)
  
      return dTdx, dtdx
  
  # Vetor espaço
  
  
  # Initial condition
  y0 = T1, t1
  sol = odeint(funct, y0 , x)
  
  # plot
  plt.plot(x, sol[:, 0], label='Trietilenoglicol')
  plt.plot(x, sol[:, 1], label='Água')
  plt.legend()
  plt.xlabel('posição')
  

  but it gives me the following error message :

  NameError                                 Traceback (most recent call last)
  <ipython-input-47-2f2a81626e49> in <module>
       35 # Initial condition
       36 y0 = T1, t1
  ---> 37 sol = odeint(funct, y0 , x)
       38 
       39 # plot
  
  c:\users\idril\appdata\local\programs\python\python36\lib\site-packages\scipy\integrate\odepack.py in odeint(func, y0, t, args, Dfun, col_deriv, full_output, ml, mu, rtol, atol, tcrit, h0, hmax, hmin, ixpr, mxstep, mxhnil, mxordn, mxords, printmessg, tfirst)
      242                              full_output, rtol, atol, tcrit, h0, hmax, hmin,
      243                              ixpr, mxstep, mxhnil, mxordn, mxords,
  --> 244                              int(bool(tfirst)))
      245     if output[-1] < 0:
      246         warning_msg = _msgs[output[-1]] + " Run with full_output = 1 to get quantitative information."
  
  <ipython-input-47-2f2a81626e49> in funct(y, x)
       23 
       24 def funct(y, x):
  ---> 25     y = T, t
       26     dTdx = (U * Di * pi / (Wt * cpt)) * (T - t)
       27     dtdx = (U * Di * pi / (Ww * cpw)) * (T - t)
  
  NameError: name 'T' is not defined
  

  I've already read many answered questions but still can't find the mistake.
• Is there a way to easily integrate a set of differential equations over a full grid of points?

  The problem is that I would like to be able to integrate the differential equations starting for each point of the grid at once instead of having to loop over the scipy integrator for each coordinate. (I'm sure there's an easy way)

  As background for the code I'm trying to solve the trajectories of a Couette flux alternating the direction of the velocity each certain period, that is a well known dynamical system that produces chaos. I don't think the rest of the code really matters as the part of the integration with scipy and my usage of the meshgrid function of numpy.

  import numpy as np
  import matplotlib.pyplot as plt
  from matplotlib.animation import FuncAnimation, writers
  from scipy.integrate import solve_ivp
  
  start_T = 100
  L = 1
  V = 1
  total_run_time = 10*3
  grid_points = 10
  
  T_list = np.arange(start_T, 1, -1)
  x = np.linspace(0, L, grid_points)
  y = np.linspace(0, L, grid_points)
  X, Y = np.meshgrid(x, y)
  condition = True
  totals = np.zeros((start_T, total_run_time, 2))
  alphas = np.zeros(start_T)
  i = 0
  
  for T in T_list:
      alphas[i] = L / (V * T)
      solution = np.array([X, Y])
      for steps in range(int(total_run_time/T)):
          t = steps*T
          if condition:
              def eq(t, x):
                  return V * np.sin(2 * np.pi * x[1] / L), 0.0
              condition = False
          else:
              def eq(t, x):
                  return 0.0, V * np.sin(2 * np.pi * x[1] / L)
              condition = True
          time_steps = np.arange(t, t + T)
          xt = solve_ivp(eq, time_steps, solution)
          solution = np.array([xt.y[0], xt.y[1]])
          totals[i][t: t + T][0] = solution[0]
          totals[i][t: t + T][1] = solution[1]
      i += 1
  
  np.save('alphas.npy', alphas)
  np.save('totals.npy', totals)
  

  The error given is :

  ValueError: y0 must be 1-dimensional.
  

  And it comes from the 'solve_ivp' function of scipy because it doesn't accept the format of the numpy function meshgrid. I know I could run some loops and get over it but I'm assuming there must be a 'good' way to do it using numpy and scipy. I accept advice for the rest of the code too.
• Handling difference in test data set and training data set

  I have collected two separate datasets one for training machine learning algorithm and other to make predictions on the test data.

  Training Dataset:

  Type of train, Date of departure, Driving safety, Driver group
  A, 2018-10-15, 5, B1
  B, 2018-11-03, 2, B1
  A, 2019-01-12, 6, A2
  

  Test Dataset:

  Type of train, Date of departure, Driving safety, Driver group
  A, 2019-05-20,, 
  B, 2018-05-21,, 
  A, 2019-06-10,, 
  C, 2019-06-14,,
  

  I am trying to find which 'Driver group' could drive the train, also notice Driving safety is not available here. As you can see 'Date of departure', and 'Type of train' differs from training to test datasets. After I have generated ARFF file to test the test dataset on trained final model, I am having issue due to above differences, but I feel these are natural differences specially Date's would differ in both train and test datasets. I am seeing following error on the weka console:

  Weka exception: training and test set are not compatible Attributes differ at position 5: Different number of labels: 89 != 250

  Given the above scenario, any suggestions how can I resolve this? Is there something I should do while training my model to be liberal on certain data elements.

  I have removed the 'Type of train' i.e., C unknown to the model. But Date of departure is something I cannot remove.

  I expect machine learning algorithm to successfully make predictions on my test data, having the said differences.

  Regards, Shravan
• How to get the log loss?

  I am playing with the Leaf Classification data set and I am struggling to compute the log loss of my model after testing it.After importing it from the metrics class here I do:

   # fitting the knn with train-test split 
   X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=42)
  
   # Optimisation via gridSearch
   knn=KNeighborsClassifier()
   params={'n_neighbors': range(1,40), 'weights':['uniform', 'distance'], 'metric':['minkowski','euclidean'],'algorithm': ['auto','ball_tree','kd_tree', 'brute']}
   k_grd=GridSearchCV(estimator=knn,param_grid=params,cv=5)
   k_grd.fit(X_train,y_train)
  
   # testing 
   yk_grd=k_grd.predict(X_test)
  
   # calculating the logloss 
   print (log_loss(y_test, yk_grd))
  

  However, my last line is resulting in the following error:

   y_true and y_pred contain different number of classes 93, 2. Please provide the true labels explicitly through the labels argument. Classes found in y_true.
  

  But the when I run the following:

  X_train.shape, X_test.shape, y_train.shape, y_test.shape, yk_grd.shape
  # results
  ((742, 192), (248, 192), (742,), (248,), (248,))
  

  what am i missing really?
• How can I use categorical and continuous variables as input to scikit logistic regression algorithm

  I am trying to perform a logistic regression on a small dataset where each instance is represented with one categorical (Male,Female) variable and one continious variable (Frequency which is between [0,1]) and dependent variable (label) is either 0 or 1. Therefore task is a binary classification problem. Below I share 5-6 examples:

  import pandas as pd 
  df = pd.read_table('dataset.csv',sep=',')
  df.head()
  
  
  Gender  Frequency  Label
  0    Male       0.10      0
  1  Female       0.23      1
  2  Female       0.35      1
  3  Female       0.21      0
  4    Male       0.15      1
  

  My question is : While performing logistic regression with scikit's logistic regression, should I encode my continious and categorical variables differently ? Does either of these variables need some special encoding ? Or is the following correct way to do that ?

  from sklearn.cross_validation import train_test_split
  X = df[['Gender','Frequency']] # Features
  y = df.Label # Target variable
  X_train,X_test,y_train,y_test=train_test_split(X,y,test_size=0.25,random_state=0)
  
  from sklearn.linear_model import LogisticRegression
  logreg = LogisticRegression()
  # fit the model with data
  logreg.fit(X_train,y_train)
  
  #
  y_pred=logreg.predict(X_test)
  

  When I do that, does the regression function understands Gender is a categorical variable and Frequency is a continuous variable ?

  EDIT: Of course, I need to convert string to float values. (E.g. "Male" -> 0 , "Female" -> 1) What I am asking is "Do I need to to something other than this ? I ask this because I saw something similar here but the thread has not been concluded therefore I couldn't be sure about validateness of answers.
• What is the fatest way to for Pose Estimation in real time?

  We are developing an pose estimation system that animate a 3D model in the screen based on the pose of the human

  The system is just single person (for now)

  We done some research about the different techniques

  1. Using multiple Web Cameras and using openPose to process
  2. Using On-hand sensors like motion band in wrist and arms
  3. Using a depth sensor like Kinct V2 / Intel RealSense / PS4 Camera / Orrbec Astra Pro

  Our Problem
  

  Our problem is , we have to decide which type gives fast real time output i.e 30fps 3D model rendering with low cost also

  Our concerns are :
  

  1. Middle end Hardware (Both devices and computer)
  2. 30FPS output

  We also concerned about the selling cost of the final product to end users

  That is , the end product selling cost must be low and in the same time it should product 30FPS output in realtime pose estimation

  Please guide me
• Rectify multiple images in Matlab

  I have 5 cameras that I like to rectify them to a common plane, I have the calibration matrices of all of them. My question is how to choose the best common plane and how to rectify all based on that?
• how to remove outlier points in PCL for region grow based surface detection

  I am using PCL to detect surface in my lidar point cloud, and I have the configuration as below:

  pcl::RegionGrowing<pcl::PointXYZ, pcl::Normal> reg;
  reg.setMinClusterSize(static_cast<int>(100));
  reg.setMaxClusterSize(static_cast<int>(1000));
  reg.setSearchMethod(new pcl::search::KdTree<pcl::PointXYZ>);
  reg.setNumberOfNeighbours(30);
  reg.setInputCloud(point_cloud);
  reg.setInputNormals(normals);
  reg.setSmoothnessThreshold(6_deg);
  reg.setCurvatureThreshold(1.0);
  

  It works fine, except I found it picks up some outlier points as shown in the image below.

  I was trying to detect a plane that comes with a leg, and it seems that the method picks up some point on the leg. Is there away to avoid such point using, e.g., density threshold in the region growing method in PCL? After spending some time on the documentation and tweaking the values for region grow setups I still could not figure out how to do it.