---
title: "movie_classification"
author: "Prof. Eric A. Suess"
format: 
  html:
    embed-resources: true
---

## Classifying newswires: A multiclass classification example

Load the data.

```{r}
library(keras)

reuters <- dataset_reuters(num_words = 10000)

c(c(train_data, train_labels), c(test_data, test_labels)) %<-% reuters
```

```{r}
str(train_data)
```

```{r}
word_index <- dataset_reuters_word_index()

reverse_word_index <- names(word_index)

names(reverse_word_index) <- as.character(word_index)

decoded_words <- train_data[[1]] %>%
  sapply(function(i) {
    if (i > 3) reverse_word_index[[as.character(i - 3)]]
    else "?"
})

decoded_review <- paste0(decoded_words, collapse = " ")

decoded_review

```

Encoding the input data

```{r}
vectorize_sequences <- function(sequences, dimension = 10000) {
  results <- matrix(0, nrow = length(sequences), ncol = dimension)
  for (i in seq_along(sequences))
    results[i, sequences[[i]]] <- 1
  results
}
```

```{r}
x_train <- vectorize_sequences(train_data)

x_test <- vectorize_sequences(test_data)
```

```{r}
to_one_hot <- function(labels, dimension = 46) {
  results <- matrix(0, nrow = length(labels), ncol = dimension)
  labels <- labels + 1

for(i in seq_along(labels)) {
  j <- labels[[i]]
  results[i, j] <- 1
  }

results

}

y_train <- to_one_hot(train_labels)

y_test <- to_one_hot(test_labels)
```

```{r}
y_train <- to_categorical(train_labels)

y_test <- to_categorical(test_labels)
```


Build the network.

```{r}
model <- keras_model_sequential() %>%
  layer_dense(64, activation = "relu") %>%
  layer_dense(64, activation = "relu") %>%
  layer_dense(46, activation = "softmax")
```

Compile the model.

```{r}
model %>% compile(optimizer = "rmsprop",
                  loss = "categorical_crossentropy",
                  metrics = "accuracy")
```


```{r}
val_indices <- 1:1000

x_val <- x_train[val_indices, ]
partial_x_train <- x_train[-val_indices, ]
y_val <- y_train[val_indices, ]
partial_y_train <- y_train[-val_indices, ]
```

```{r}
history <- model %>% fit(partial_x_train, 
                         partial_y_train, 
                         epochs = 20, 
                         batch_size = 512, 
                         validation_data = list(x_val, y_val))
```

```{r}
plot(history)
```

# Retrain the data.

```{r}
model <- keras_model_sequential() %>% 
  layer_dense(64, activation = "relu") %>%
  layer_dense(64, activation = "relu") %>%
  layer_dense(46, activation = "softmax")

model %>% compile(optimizer = "rmsprop",
                  loss = "categorical_crossentropy",
                  metrics = "accuracy")

model %>% fit(x_train, y_train, epochs = 9, batch_size = 512)

results <- model %>% evaluate(x_test, y_test)
results
```

```{r}
mean(test_labels == sample(test_labels))
```

Predictions.  Inferences.

```{r}
predictions <- model %>% predict(x_test)
predictions
```

```{r}
str(predictions)

sum(predictions[1, ])

which.max(predictions[1, ])
```


A different way.

```{r}
y_train <- train_labels

y_test <- test_labels
```

```{r}
model %>% compile(
  optimizer = "rmsprop",
  loss = "sparse_categorical_crossentropy",
  metrics = "accuracy")
```

```{r}
model <- keras_model_sequential() %>%
  layer_dense(64, activation = "relu") %>%
  layer_dense(4, activation = "relu") %>%
  layer_dense(46, activation = "softmax")

model %>% compile(optimizer = "rmsprop",
                  loss = "categorical_crossentropy",
                  metrics = "accuracy")

model %>% fit(partial_x_train, partial_y_train,  
              epochs = 20, batch_size = 128, validation_data = list(x_val, y_val))
```