metric_pearson_correlation {ttgsea} | R Documentation |
Pearson correlation coefficient can be seen as one of the model performance metrics. This is a measure of how close the predicted value is to the true value. If it is close to 1, the model is considered a good fit. If it is close to 0, the model is not good. A value of 0 corresponds to a random prediction.
Dongmin Jung
keras::k_mean, keras::sum, keras::k_square, keras::k_sqrt
library(reticulate) if (keras::is_keras_available() & reticulate::py_available()) { num_tokens <- 1000 length_token <- 30 embedding_dims <- 50 num_units_1 <- 32 num_units_2 <- 16 stacked_gru <- function(num_tokens, embedding_dims, length_seq, num_units_1, num_units_2) { model <- keras::keras_model_sequential() %>% keras::layer_embedding(input_dim = num_tokens, output_dim = embedding_dims, input_length = length_seq) %>% keras::layer_gru(units = num_units_1, activation = "relu", return_sequences = TRUE) %>% keras::layer_gru(units = num_units_2, activation = "relu") %>% keras::layer_dense(1) model %>% keras::compile(loss = "mean_squared_error", optimizer = "adam", metrics = custom_metric("pearson_correlation", metric_pearson_correlation)) } }