Trees and Forests Assignment
Anvesh Raavi
2020-09-21
Load the libraries + functions
library(Rling)
library(party)
library(dplyr)
library(skimr)
library(DataExplorer)
data(nerd)
head(nerd)## Noun Num Century Register Eval
## 1 nerd pl XX ACAD Neutral
## 2 nerd pl XXI ACAD Neutral
## 3 nerd pl XX ACAD Neutral
## 4 nerd pl XX ACAD Neutral
## 5 nerd pl XX ACAD Neutral
## 6 nerd pl XXI ACAD NeutralDescription of the data
Dependent variable:
- Noun: which category is represented either nerd or geek.
Independent variables:
- Num: a measure of social group, either pl (plural) or sg (single)
- Century: when the word was used, as XX (20th) or XXI (21st) century
- Register: information about the context of the word; coded as ACAD (academic), MAG (magazine), NEWS (newspapers), and SPOK (spoken)
- Eval: A measure of the valence of the word, Neg for negative, Neutral, and Positive
summary(nerd)## Noun Num Century Register Eval
## geek:670 pl:436 XX :515 ACAD: 54 Neg :132
## nerd:646 sg:880 XXI:801 MAG :622 Neutral :896
## NEWS:343 Positive:288
## SPOK:297dim(nerd)## [1] 1316 5glimpse(nerd)## Rows: 1,316
## Columns: 5
## $ Noun <fct> nerd, nerd, nerd, nerd, nerd, nerd, nerd, nerd, nerd, nerd...
## $ Num <fct> pl, pl, pl, pl, pl, pl, pl, pl, pl, pl, pl, pl, pl, pl, pl...
## $ Century <fct> XX, XXI, XX, XX, XX, XXI, XXI, XX, XXI, XX, XXI, XX, XXI, ...
## $ Register <fct> ACAD, ACAD, ACAD, ACAD, ACAD, ACAD, ACAD, ACAD, ACAD, ACAD...
## $ Eval <fct> Neutral, Neutral, Neutral, Neutral, Neutral, Neutral, Neut...skim(nerd)| Name | nerd |
| Number of rows | 1316 |
| Number of columns | 5 |
| _______________________ | |
| Column type frequency: | |
| factor | 5 |
| ________________________ | |
| Group variables | None |
Variable type: factor
| skim_variable | n_missing | complete_rate | ordered | n_unique | top_counts |
|---|---|---|---|---|---|
| Noun | 0 | 1 | FALSE | 2 | gee: 670, ner: 646 |
| Num | 0 | 1 | FALSE | 2 | sg: 880, pl: 436 |
| Century | 0 | 1 | FALSE | 2 | XXI: 801, XX: 515 |
| Register | 0 | 1 | FALSE | 4 | MAG: 622, NEW: 343, SPO: 297, ACA: 54 |
| Eval | 0 | 1 | FALSE | 3 | Neu: 896, Pos: 288, Neg: 132 |
Conditional inference model
set.seed(123456)
cimodel = ctree(Noun ~ Num + Century + Register + Eval, data = nerd)Make a plot
- Plot the conditional inference model.
plot(cimodel)
Interpret the categories
Write out an interpretation of the results from the model. You can interpret the branches of the tree to determine what features defines each category.
- First split is using Eval levels
- The next split is using century
- In node3, there are 194 observations, about 10% of those are nerd and the rest is geek. Similar interpretations can be done on other nodes
- For twentieth century there are more negative or neutral observations of the word nerd.
Conditional inference model predictiveness
predictiveness = table(predict(cimodel),nerd$Noun)
predictiveness##
## geek nerd
## geek 227 61
## nerd 443 585sum(diag(predictiveness))/sum(predictiveness) * 100## [1] 61.70213Random forests
forest = cforest(Noun ~ Num + Century + Register + Eval, data = nerd, controls = cforest_unbiased(ntree = 1000, mtry = 2))Variable importance
Which variables were the most important?
- Eval and Century are most important variables
variable_importance = varimp(forest,conditional = T)
round(variable_importance, 3)## Num Century Register Eval
## -0.002 0.024 -0.002 0.055dotchart(sort(variable_importance),main = "Importance of variables")
Forest model predictiveness
Did it do better than the conditional inference tree?
- Random forest model is slightly better than the tree model
predictions = table(predict(forest),nerd$Noun)
predictions##
## geek nerd
## geek 325 136
## nerd 345 510sum(diag(predictions))/sum(predictions) * 100## [1] 63.44985library(reticulate)
py_config()## python: C:/Users/raavi/AppData/Local/r-miniconda/envs/r-reticulate/python.exe
## libpython: C:/Users/raavi/AppData/Local/r-miniconda/envs/r-reticulate/python36.dll
## pythonhome: C:/Users/raavi/AppData/Local/r-miniconda/envs/r-reticulate
## version: 3.6.11 (default, Aug 5 2020, 19:41:03) [MSC v.1916 64 bit (AMD64)]
## Architecture: 64bit
## numpy: C:/Users/raavi/AppData/Local/r-miniconda/envs/r-reticulate/Lib/site-packages/numpy
## numpy_version: 1.19.1py_module_available("sklearn")## [1] TRUEPython model
import sklearn
from sklearn import tree
import pandas as pd
data = pd.get_dummies(r.nerd[["Num","Century","Register","Eval"]])
data.head()## Num_pl Num_sg Century_XX ... Eval_Neg Eval_Neutral Eval_Positive
## 0 1 0 1 ... 0 1 0
## 1 1 0 0 ... 0 1 0
## 2 1 0 1 ... 0 1 0
## 3 1 0 1 ... 0 1 0
## 4 1 0 1 ... 0 1 0
##
## [5 rows x 11 columns]nouns = pd.get_dummies(r.nerd["Noun"])
nouns.head()## geek nerd
## 0 0 1
## 1 0 1
## 2 0 1
## 3 0 1
## 4 0 1Create the Tree
decision_tree = tree.DecisionTreeClassifier()
decision_tree = decision_tree.fit(data,nouns)Printing out the Tree
print(tree.export_text(decision_tree,feature_names = list(data.columns)))## |--- Eval_Positive <= 0.50
## | |--- Century_XXI <= 0.50
## | | |--- Num_pl <= 0.50
## | | | |--- Register_ACAD <= 0.50
## | | | | |--- Register_MAG <= 0.50
## | | | | | |--- Register_SPOK <= 0.50
## | | | | | | |--- Eval_Neg <= 0.50
## | | | | | | | |--- class: 0
## | | | | | | |--- Eval_Neg > 0.50
## | | | | | | | |--- class: 0
## | | | | | |--- Register_SPOK > 0.50
## | | | | | | |--- Eval_Neg <= 0.50
## | | | | | | | |--- class: 0
## | | | | | | |--- Eval_Neg > 0.50
## | | | | | | | |--- class: 1
## | | | | |--- Register_MAG > 0.50
## | | | | | |--- Eval_Neg <= 0.50
## | | | | | | |--- class: 0
## | | | | | |--- Eval_Neg > 0.50
## | | | | | | |--- class: 0
## | | | |--- Register_ACAD > 0.50
## | | | | |--- Eval_Neutral <= 0.50
## | | | | | |--- class: 0
## | | | | |--- Eval_Neutral > 0.50
## | | | | | |--- class: 0
## | | |--- Num_pl > 0.50
## | | | |--- Register_ACAD <= 0.50
## | | | | |--- Register_MAG <= 0.50
## | | | | | |--- Eval_Neutral <= 0.50
## | | | | | | |--- Register_NEWS <= 0.50
## | | | | | | | |--- class: 0
## | | | | | | |--- Register_NEWS > 0.50
## | | | | | | | |--- class: 0
## | | | | | |--- Eval_Neutral > 0.50
## | | | | | | |--- Register_SPOK <= 0.50
## | | | | | | | |--- class: 0
## | | | | | | |--- Register_SPOK > 0.50
## | | | | | | | |--- class: 0
## | | | | |--- Register_MAG > 0.50
## | | | | | |--- Eval_Neutral <= 0.50
## | | | | | | |--- class: 0
## | | | | | |--- Eval_Neutral > 0.50
## | | | | | | |--- class: 0
## | | | |--- Register_ACAD > 0.50
## | | | | |--- class: 0
## | |--- Century_XXI > 0.50
## | | |--- Num_pl <= 0.50
## | | | |--- Register_ACAD <= 0.50
## | | | | |--- Register_MAG <= 0.50
## | | | | | |--- Eval_Neg <= 0.50
## | | | | | | |--- Register_SPOK <= 0.50
## | | | | | | | |--- class: 0
## | | | | | | |--- Register_SPOK > 0.50
## | | | | | | | |--- class: 1
## | | | | | |--- Eval_Neg > 0.50
## | | | | | | |--- Register_SPOK <= 0.50
## | | | | | | | |--- class: 1
## | | | | | | |--- Register_SPOK > 0.50
## | | | | | | | |--- class: 0
## | | | | |--- Register_MAG > 0.50
## | | | | | |--- Eval_Neutral <= 0.50
## | | | | | | |--- class: 0
## | | | | | |--- Eval_Neutral > 0.50
## | | | | | | |--- class: 1
## | | | |--- Register_ACAD > 0.50
## | | | | |--- Eval_Neutral <= 0.50
## | | | | | |--- class: 1
## | | | | |--- Eval_Neutral > 0.50
## | | | | | |--- class: 0
## | | |--- Num_pl > 0.50
## | | | |--- Register_MAG <= 0.50
## | | | | |--- Eval_Neutral <= 0.50
## | | | | | |--- Register_SPOK <= 0.50
## | | | | | | |--- Register_ACAD <= 0.50
## | | | | | | | |--- class: 0
## | | | | | | |--- Register_ACAD > 0.50
## | | | | | | | |--- class: 0
## | | | | | |--- Register_SPOK > 0.50
## | | | | | | |--- class: 0
## | | | | |--- Eval_Neutral > 0.50
## | | | | | |--- Register_ACAD <= 0.50
## | | | | | | |--- Register_NEWS <= 0.50
## | | | | | | | |--- class: 0
## | | | | | | |--- Register_NEWS > 0.50
## | | | | | | | |--- class: 0
## | | | | | |--- Register_ACAD > 0.50
## | | | | | | |--- class: 0
## | | | |--- Register_MAG > 0.50
## | | | | |--- Eval_Neutral <= 0.50
## | | | | | |--- class: 1
## | | | | |--- Eval_Neutral > 0.50
## | | | | | |--- class: 0
## |--- Eval_Positive > 0.50
## | |--- Century_XXI <= 0.50
## | | |--- Register_MAG <= 0.50
## | | | |--- Register_ACAD <= 0.50
## | | | | |--- Num_sg <= 0.50
## | | | | | |--- Register_SPOK <= 0.50
## | | | | | | |--- class: 0
## | | | | | |--- Register_SPOK > 0.50
## | | | | | | |--- class: 0
## | | | | |--- Num_sg > 0.50
## | | | | | |--- Register_SPOK <= 0.50
## | | | | | | |--- class: 0
## | | | | | |--- Register_SPOK > 0.50
## | | | | | | |--- class: 0
## | | | |--- Register_ACAD > 0.50
## | | | | |--- class: 0
## | | |--- Register_MAG > 0.50
## | | | |--- Num_sg <= 0.50
## | | | | |--- class: 1
## | | | |--- Num_sg > 0.50
## | | | | |--- class: 1
## | |--- Century_XXI > 0.50
## | | |--- Register_MAG <= 0.50
## | | | |--- Register_ACAD <= 0.50
## | | | | |--- Register_SPOK <= 0.50
## | | | | | |--- Num_sg <= 0.50
## | | | | | | |--- class: 1
## | | | | | |--- Num_sg > 0.50
## | | | | | | |--- class: 1
## | | | | |--- Register_SPOK > 0.50
## | | | | | |--- Num_pl <= 0.50
## | | | | | | |--- class: 1
## | | | | | |--- Num_pl > 0.50
## | | | | | | |--- class: 1
## | | | |--- Register_ACAD > 0.50
## | | | | |--- class: 1
## | | |--- Register_MAG > 0.50
## | | | |--- Num_sg <= 0.50
## | | | | |--- class: 1
## | | | |--- Num_sg > 0.50
## | | | | |--- class: 1Confusion Matrix
prediction = pd.DataFrame(decision_tree.predict(data))
prediction.columns = list(nouns.columns)
prediction_category = prediction.idxmax(axis = 1)
variable_category = nouns.idxmax(axis = 1)
sklearn.metrics.confusion_matrix(prediction_category,variable_category,labels=["nerd","geek"])## array([[439, 269],
## [207, 401]], dtype=int64)Thought questions
Are the models easier to create using R or Python (your own thoughts, they can be different than what I said in the lecture)?
- Easier to create models using R as the visualizations are inbuilt and python requires creation of dummy variables
Which model gave you a better classification of the categories?
- Random forest model
What other variables might be useful in understanding the category membership of geek versus nerd? Basically, what could we add to the model to improve it (there’s no one right answer here - it’s helpful to think about what other facets we have not measured)?
- Gender, age, locations would be helpful to get more insights. An explanation of Eval column would be helpful to understand the dataset context more.