Data Science Projects

Dimension Reduction	Clustering
Classification methods (Logistic Regression, Decision Tree, Random Forest) after applying PCA wine quality classification Data Set: Wine Quality Dataset Applied PCA for dimension reduction on input vectors Algorithms: Logistic Regression, Decision Tree, Random Forest (using Scikit-learn ) Compared performance of algorithms	K-Means (using Scikit-learn ) Discover classes in a given data set Data Set: Wine Quality Dataset Algorithm: K-Means clustering (using Scikit-learn)
Regression	Classification
Regression methods (Linear, Ridge, Lasso, ElasticNet, Huber) Predict price of a house Data Set: House Sales in King County, USA Algorithms: Linear, Ridge, Lasso, ElasticNet, Huber (using Scikit-learn)	DNNClassifier, LinearClassifier (using TensorFlow) Classifiy wine quality based on given measurements Data set: Wine Quality Dataset Classification Algorithm: DNNClassifier (multi-layer neural network) Compare performans of classification algorithms
Visualization (Exploratory Data Analysis - EDA)
Matplotlib	Seaborn
Matplotlib Example	Seaborn Example
Convolutional-NN	Recurrent-NN
Autoencoders	Transformers
Reinforcement Learning	GAN

# Data Science and Machine Learning Notebooks

• Visualization for Exploratory Data Analysis
  • Matplotlib notebook (scatter plot, histogram/kde, box plot)

  • Seaborn notebook

  • Lightning - http://lightning-viz.org/
  • Bokeh - https://bokeh.pydata.org/
• Data cleaning
  • Notebook
• Dimension reduction
  • Notebook for PCA, LogisticRegression, DecisionTree, RandomForest (from Scikit-learn )
    • Data Set: Wine Quality
    • Problem: Predict quality of wine.
    • Approach: Used classification methods (Logistic Regression, Decision Tree, Random Forest) after applying PCA.
  • Notebook for PCA, LDA, Autoencoder (from Scikit-learn )
    • Data Set: Human Activity Recognition
    • Problem: Predict activity from observations.
    • Approach:

• Clustering

  • Notebook for K-Means (from Scikit-learn )
    • Data Set: Wine Quality
    • Problem: Cluster data for two quality categories.
    • Approach: Used K means algorithm to cluster.

• Regression

  • Notebook for LinearRegression, Ridge, Lasso, ElasticNet, ElasticNet CrossValidation, HuberRegressor (from Scikit-learn )
    • Data Set: House Sales in King County, USA
    • Problem: Predict price of a house.
    • Approach: Used regression methods (Linear, Ridge, Lasso, ElasticNet, Huber).
  • TensorFlow.LinearRegressor
  • TensorFlow.DNNRegressor

• Classification

  • Notebook for DNNClassifier, LinearClassifier (from TensorFlow)
    • Data Set: Wine Quality
    • Problem: Predict quality of wine.
    • Approach: Used classification methods (DNNClassifier, LinearClassifier).
  • SVM
• Text Analysis and NLP
  • Word2Vec
• Reinforcement Learning