Kaggle Competition Practice

Housing Prices Competition

Kaggle Housing Prices Competition Link

Objective: Predict the sales price of individual residential property in Ames, Iowa from 2006 to 2010. For each Id in the test set, a prediction value should be populated for the SalePrice variable.

Notes on Submissions:

• first-submission.csv

  • Algorithm: Random Forests
  • Features: ['LotArea', 'YearBuilt', '1stFlrSF', '2ndFlrSF', 'FullBath', 'BedroomAbvGr', 'TotRmsAbvGrd']
  • Root-Mean-Squared-Error (RMSE) score: 22337.06

• second-submission.csv

  • Algorithm: Gradient Boosting Regression
  • Features: All table variables
  • Used Hyperparameter Tuning - GridSearchCV
  • Root-Mean-Squared-Error (RMSE) score: 182906.48

• third-submission.csv

  • Algorithm: Gradient Boosting Regression
  • Features: ['LotArea', 'YearBuilt', '1stFlrSF', '2ndFlrSF', 'FullBath', 'BedroomAbvGr', 'TotRmsAbvGrd']
  • Used Hyperparameter Tuning - GridSearchCV
  • Root-Mean-Squared-Error (RMSE) score: 182906.48

• fourth-submission.csv

  • Algorithm: Linear Regression
  • Features: ['MSSubClass', 'LotFrontage', 'LotArea', 'OverallQual', 'OverallCond', 'YearBuilt', 'YearRemodAdd', 'MasVnrArea', 'BsmtFinSF1', 'BsmtUnfSF', 'TotalBsmtSF', '1stFlrSF', '2ndFlrSF', 'LowQualFinSF', 'GrLivArea', 'FullBath', 'HalfBath', 'BedroomAbvGr', 'KitchenAbvGr', 'TotRmsAbvGrd', 'Fireplaces', 'GarageCars', 'WoodDeckSF', 'OpenPorchSF', 'EnclosedPorch', 'ScreenPorch', 'PoolArea']
  • Root-Mean-Squared-Error (RMSE) score: 20476.40

---

Titanic Competition

Kaggle Titanic Competition Link

Objective: Use machine learning to create a model that predicts which passengers survived the Titanic shipwreck by using passenger data (ie name, age, gender, socio-economic class, etc).

Notes on Submissions:

• first-submission.csv

  • Algorithm: Random Forests
  • Features: ['Pclass', 'Sex', 'SibSp', 'Parch']
  • Categorization accuracy score: 0.59569

• second-submission.csv

  • Algorithm: Random Forests
  • Features: ['Pclass', 'Sex', 'SibSp', 'Parch']
  • Used Hyperparameter Tuning - GridSearchCV
  • Categorization accuracy score: 0.76555

• third-submission.csv

  • Algorithm: Logistic Regression
  • Features: ['Pclass', 'Sex', 'SibSp', 'Parch']
  • Used Hyperparameter Tuning - GridSearchCV
  • Categorization accuracy score: 0.77511

• fourth-submission.csv

  • Algorithm: Random Forests
  • Features: ['Pclass','Sex','Age','Fare','Family_cnt','Cabin_ind']
  • Used Hyperparameter Tuning - GridSearchCV
  • Categorization accuracy score: 0.78708