• matplotlib
• scipy
• seaborn
• pandas
• numpy
• scikit-learn
• tensorflow
• pyodc (if you need to connect to a database)

For this project I used a connection to a SQL Database, so the pyodbc library is installed. But you could either import a csv file or make a connection to where your dataset is located. The 'database' folder contains some csv files used to explain some classes like Regressions and Clusters

1. Pareto's Chart
2. Histogram
3. Bars' charts
4. Scatter Plot
5. Probability Plot

6. Skew
7. Variance, Covariance, Standard Deviation and Variance coefficient
8. Correlation coefficient
9. Standarize Normal Distribution (zscore)
10. Central Limit Theorem
11. Confidence Intervals. Z & T for var population known or unknown, respectively
12. Confidence Interval. Dependent Samples
13. Confidence Interval. Independent Samples
    1. Var populations known (populations normally distributed, var populations known, sample sizes differ)
    2. Var populations unknown but assumed are equals
    3. Var populations unknown but assumed are different
14. Hypothesis Test. (Applied for each type of Confidence Interval)

1. Simple linear Regression (OLS)
2. Multiple linear Regression (OLS)
   1. Dummy variables
3. Logistic Regression
   1. Binary predictors

1. Supervised Models
   1. Regressions
   2. Classification
2. Non-supervised Models (PENDING)
3. Objective Function
   1. Loss Norma-L2 (Regressions)
   2. Cross Entropy (Classification)
4. Optimization Algorithm
   1. Gradient descent

1. Probability distributions
2. Plot Confidence Intervals in normal distributions and probplots
3. Simple linear Regression
   1. Other solution methods
   2. Log transformation variables as elasticity (economics)
   3. Durbin-Watson to detect no autocorrelation
   4. Other regression models to time-series or when the error terms (No Autocorrelation) are correlated
4. Hiperparameters vs Parameters
   1. Amplitud
   2. Deep
   3. Learning rate (eta)
   4. Lot size
   5. Momentum coefficient (alpha)
   6. Decay coefficient (c)
5. DeepLearning Activation functions
   1. sigmoide (0, 1)
   2. tanh (-1, 1)
   3. ReLu (0, inf)
   4. softmax (0, 1)
6. Preventing overfitting. Early stop
   1. Set number of epochs
   2. Stop when delta <= 0.001
   3. Validation / Cross-validation
7. Initialization
8. Find the Global minimum (not local)
9. Programming the Learning rate
10. Preprocessing
    1. Standarization
    2. Normalization
    3. PCA
    4. Bleaching
11. Categorical data
    1. Binary
    2. One-Hot
12. Balanced previous/data
13. RNCs and RNRs
14. Other Deep Learning models
    1. Bayesian nets
    2. Generative models
    3. Decision Trees