This is an attempt to use Deep Learning (LSTM) to predict where to cut a Thai word given a Thai String.
For example, suppose "ไก่จิกเด็กตายบนปากโอ่ง" was given, the tokenizer should seperate the string into "ไก่|จิก|เด็ก|ตาย|บน|ปาก|โอ่ง"
Result
Test set 1 Start comparing Accuracy : 96.76 Precision : 90.18 Recall : 94.02 F1 Score : 92.06
Test set 2 Start comparing Accuracy : 96.51 Precision : 90.35 Recall : 92.61 F1 Score : 91.47
Test set 3 Start comparing Accuracy : 96.50 Precision : 90.79 Recall : 92.27 F1 Score : 91.52
Test set 4 Start comparing Accuracy : 96.08 Precision : 89.24 Recall : 91.87 F1 Score : 90.53