3catz

#!/usr/bin/env python
# encoding: utf-8

import tweepy #https://github.com/tweepy/tweepy
import csv

#Twitter API credentials
consumer_key = ""
consumer_secret = ""
access_key = ""

3catz

3catz

trainstock = yf.Ticker("SPY")
start = "2009-01-01"
end = "2016-01-01"

st = trainstock.history(start = start,end = end)
st = st[["Close","Open","Volume","High","Low"]]

D = reconstruct(st["Close"].values, dim = 45, tau = 1)
win = D[:,:-1] ; s = D[:,-1]
std = np.std(win, axis = -1)

3catz

from bayes_opt import BayesianOptimization

def RF_opt(n_estimators, max_depth): 
    
  global rskf 
  reg = RandomForestClassifier(verbose = 0,
                          n_estimators = int(n_estimators),
                          #min_samples_split = int(min_samples_split),
                          #min_samples_leaf = int(min_samples_leaf),
                          max_depth = int(max_depth),

3catz

oversampler = MulticlassOversampling(sv.TRIM_SMOTE(proportion = 0.1))
warnings.filterwarnings("ignore")
Scores1 = []
cmatrices1 = []
cmatrices2 = []
Scores2 = [] 
for i in range(50):
  print("Trial {}".format(i))
  print("-----------------------------")
  scores1 = [] 

3catz

def getWeights(d,lags):
    # return the weights from the series expansion of the differencing operator
    # for real orders d and up to lags coefficients
    w=[1]
    for k in range(1,lags):
        w.append(-w[-1]*((d-k+1))/k)
    w=np.array(w).reshape(-1,1) 
    return w

def plotWeights(dRange, lags, numberPlots):

3catz

covariates = trainx; target = trainy
def lgb_trainer(num_leaves, learning_rate, 
                max_depth, n_estimators, 
                reg_lambda,
                #alpha, 
                reg_alpha, 
                subsample):
  
  lgb = LGBMRegressor(objective = "quantile",
                      alpha = .95,

3catz

scores = [] 
for i in range(len(report)):
  
  myfeats = report.iloc[i,1] ; print(myfeats)
  X = D[myfeats] ; y = y
  clf = LogisticRegression(solver = "liblinear", C = 6, tol = 1)
  #clf = RandomForestClassifier()
  rskf = RepeatedStratifiedKFold(n_splits = 10, n_repeats = 100)
  score = np.mean(cross_val_score(clf, X, y, cv = rskf, scoring = "roc_auc"))
  scores.append(score) 

3catz

from sklearn.feature_selection import *
feat_list = [] 
all_scores = [] 
for i in range(10):
  np.random.seed(i)
  sfm = SelectFromModel(estimator = clf, threshold=None, prefit=False, 
                       norm_order=1, max_features = 12)
  sfm.fit(D[allfeats], y)
  modfeats = sfm.get_support()
  

3catz

from pycaret.datasets import get_data
from pycaret.classification import *
report["Scores"] = np.round(report["Scores"], 3)
report.sort_values(by = "Scores", ascending = False, inplace = True)
#report.index
ga_feats = report.iloc[0]["Chosen Feats"]
ename = setup(data = D[used_feats], target = "DEATH_EVENT", 
              test_data = None,
              fold_strategy = "stratifiedkfold",
              fold_shuffle = True,