groverpr

ratings = pd.read_csv('ratings_small.csv') # loading data from csv
"""
ratings_small.csv has 4 columns - userId, movieId, ratings, and timestammp
it is most generic data format for CF related data
"""

val_indx = get_cv_idxs(len(ratings))  # index for validation set
wd = 2e-4 # weight decay
n_factors = 50 # n_factors - dimension of embedding matrix (D)

groverpr

# Split data into 5 folds

data.split(n_folds=5)

from surprise import SVD, evaluate
from surprise import NMF

# svd
algo = SVD()
evaluate(algo, data, measures=['RMSE'])

groverpr

# http://surprise.readthedocs.io/en/stable/getting_started.html
# I believe in loading all the datasets from pandas df 
# you can also load dataset from csv and whatever suits

ratings = pd.read_csv('ratings_small.csv') # reading data in pandas df

from surprise import Reader, Dataset

# to load dataset from pandas df, we need `load_fromm_df` method in surprise lib

groverpr

    # plotting after prediction
    xa = np.array(x.x) # column name of x is x 
    order = np.argsort(xa)
    xs = np.array(xa)[order]
    ys = np.array(predf)[order]
    
    #epreds = np.array(epred[:,None])[order]

    f, (ax1, ax2) = plt.subplots(1, 2, sharey=True, figsize = (13,2.5))

groverpr

    # predictions by ith decisision tree
    
    predi = np.zeros(n)
    np.put(predi, left_idx, np.repeat(np.mean(yi[left_idx]), r))  # replace left side mean y
    np.put(predi, right_idx, np.repeat(np.mean(yi[right_idx]), n-r))  # right side mean y
    
    predi = predi[:,None]  # make long vector (nx1) in compatible with y
    predf = predf + predi  # final prediction will be previous prediction value + new prediction of residual
    
    ei = y - predf  # needed originl y here as residual always from original y    

groverpr

xi = x # initialization of input
yi = y # initialization of target
# x,y --> use where no need to change original y
ei = 0 # initialization of error
n = len(yi)  # number of rows
predf = 0 # initial prediction 0

for i in range(30): # loop will make 30 trees (n_estimators). 
    tree = DecisionTree(xi,yi) # DecisionTree scratch code can be found in shared github/kaggle link. 
                               # It just create a single decision tree with provided min. sample leaf

groverpr

x = np.arange(0,50)
x = pd.DataFrame({'x':x})

# just random uniform distributions in differnt range

y1 = np.random.uniform(10,15,10)
y2 = np.random.uniform(20,25,10)
y3 = np.random.uniform(0,5,10)
y4 = np.random.uniform(30,32,10)
y5 = np.random.uniform(13,17,10)

groverpr

from waterfallcharts import quick_charts as qc
a = [‘Bias’, ‘Age’, ‘Sex’, ‘Blood Pressure’]
b = [0.3, 0.6, -0.1, -0.2]
plot = qc.waterfall(a,b, Title= ‘Patient A’, y_lab= ‘Predicted probability’, x_lab= ‘Contributing features (path)’,
 net_label = ‘Final Prediction’)
plot.show()

groverpr

#Pseudo code: 
def pred_ci(model, x_val, percentile = 95, n_pnt):   
    
    """
    x_val = validation input
    percentile = required confidence level
    model = random forest model
    """

    allTree_preds = np.stack([t.predict(x_val) for t in model.estimators_], axis = 0)

groverpr

# defining rmse as scoring criteria (any other criteria can be used in a similar manner)

def score(x1,x2):
    return metrics.mean_squared_error(x1,x2)
# defining feature importance function based on above logic
def feat_imp(m, x, y, small_good = True): 
"""
m: random forest model
x: matrix of independent variables
y: output variable