This repo is the summary for the work of Melbourne house price prediction.
Under report folder, contain source code for reporting.
report/README.Rmd: R markdown to provide data insight of Melbourne house price data.
Further report detail under report folder.
Under process folder, contain source code for data processing.
process/spilt_train_data.R: R Script to split training data at local train set and local test set under 85/15 ratio.
Further processing detail under process folder.
Under model folder, contain source code of models to predict Melbourne house price:
-
model/model_decision_tree.R: R Script of Decision Tree Model. -
model/model_random_forest.R: R Script of Random Forest Model. -
model/model_gbm.R: R Script of Gradient Boosting Model.
Further modeling detail under model folder.
Data source from Monash University, access the data from this competition vitticeps in Kaggle.
14 Features in this dataset:
-
id: unique id for property. -
suburb: suburb location of property. -
result:Sindicates property sold;SP- property sold prior;PI- property passed in;PN- sold prior not disclosed;SN- sold not disclosed;NB- no bid;VB- vendor bid;o res- other residential;w- withdrawn prior to auction. -
rating -
nbeds: number of bedrooms. -
property_type:h= house,t= townhouse,u= unit/apartment. -
day: day of the month of auction. -
month: month of auction. -
year: year of auction. -
nvisits: How many people came to open houses. -
ncars: Number of parking places. -
nbaths: Number of bathrooms. -
land_size: Size of the lot, in sq m, units will be 0. -
house_size: Internal size of property in sq m.