Knn_classification.Rmd

---
title: "Project_2"
author: "4 Idiots Group"
date: "December 4, 2018"
output:
  word_document: default
  html_document: default
---

```{r}
require(tidyverse)
require(class)

```

```{r}

nyc_data = read_csv('NYC_Jobs_2.csv')
nyc_data

```

```{r}
select_data = select(nyc_data, c(`Work Location`, IT, Non_IT))

salary = round((nyc_data$Annual_salary_from +  nyc_data$Annual_Salary_to) /2)


new_data = mutate(select_data, salary = salary)

location_openings =gather(new_data, IT_cat, Openings, 2:3)

category =as.numeric(as.factor(location_openings$IT_cat))

final_data_0 = mutate(location_openings, IT=category )
final_data = filter(final_data_0,Openings != 0)
final_data

```


```{r}

input = subset(final_data, select = c(salary, Openings))
label = final_data$IT_cat

```

```{r}
input_n = sapply(input, function(x){(x-min(x))/(max(x)-min(x))}) 

```

```{r}

location_dummies = model.matrix(~`Work Location`-1,data=final_data) 

input_n_new = data.frame(input_n, location_dummies)
#input_n_new

```

```{r}
set.seed(1234)
indices = sample(1:2, size=nrow(input_n_new), replace = T, prob = c(.8,.2))


data = data.frame(indices==1, input_n_new)

training_input = input_n_new[indices == 1,]
testing_input = input_n_new[indices == 2,]

training_label = label[indices==1]
testing_label = label[indices==2]

```

```{r}


set.seed(1234)

#sqrt(nrow(training_input))

predications = knn(train = training_input, test=testing_input,cl=training_label, k=13)

sum(predications==testing_label)/length(testing_label)

table(predications,testing_label)


```