Data preprocessing

• The first step：

python fasta2word.py -fasta xxx.fasta

• The second step(XXX is the first step to get the word segmentation file, which will eventually generate a word vector model of word2vec.model)

python word2vec.py -word xxx

Parameter Settings

---

parameter	values
-train_data_path	train file
-train_pos	Number of positive examples
-train_neg	Number of negative examples
-seed	seed
-freeze	Embedding freeze
-embedding1	Word embedding model
-batch_size	batch size
-test_data_path	test file
-test_pos	Number of positive examples
-test_neg	Number of negative examples
-rnn_layers	lstm layers
-fix_len	fix length
-learning_rate	learning rate
-dropout	dropout
-hidden_dims	hidden_dim of lstm
-num_epochs	epochs
-init	default：True

---

Example

• Independent testing（When no test set is provided, the training set is partitioned）

python start_train.py
-train_data_path
train_data.txt
-train_pos
154000
-train_neg
154000
-embedding1
word2vec.model
-fix_len
39
# Nonessential parameter
-test_data_path
test_data.txt
-test_pos
880
-test_neg
880