We used a few simulated dataset to demonstrate the nuts and bolts of a few trees and rules machine learning algorithm, random forest, cforest, gbm, cubist. We also looked into details the issues of correlated variables' impact on model, the impact of different bagging and learning rates, and the variable granularity.
Using the chemical manufacturing process database (K&J 7.5), applies a few tree based system to decide the optimal model and compare them
This project uses data imputation, data splitting, and pre-processing steps , and train several nonlinear regression models, on the K&J 7.5 chemical manufacturing data set. The models used included KNN, SVM, MARS, NN. Models performance were done.
Friedman (1991) introduced several benchmark data sets create by simulation. This document uses a few models to evaluate it (KNN, MARS,SVM,MARS)
PLS model tuning is used to determine the permeability of potential drug compounds.
Using the data of chemical manufacturing process, the objective is to understand the relationship between biological measurements of the raw materials (predictors), measurements of the manufacturing process (predictors), and the response of product yield (outcome).
ARIMA Model, HA8.1-8.7
Research Question:How can I recommend the top 10 movies to certain people (users), based on their history of preference, or on the experience of similar people like them, or some other mechanism?Q Content based filtering, User based filtering, and SVD.
Content based filtering, User based filtering, and SVD are performed on the movielense data within the recommenderlab package. Their performance is also compared.
This project compares different recommendation systems for certain users, on the jokes they might like. The comparisons are based on users (UBCF) and items (IBCF). Different normalization mechanisms are also compared.
This is the singular value decomposition methods for the recommendation system, using movie lense dataset
Read Json, HTML, XML data into R
Hui (Gracie) Han Solution