DATASETS

This study proposes a hybrid model: LSHS integrating Long Short-Term Memory (LSTM) and the Harmony Search optimization method to enhance the accuracy of software effort estimation. To evaluate its effectiveness, we evaluate our proposed model on four diverse datasets — China, Finnish, Desharnais, and Albrecht — with the number of records of 499, 407, 81, and 24, respectively. Our model is systematically benchmarked against eight existing machine learning models, employing 3 evaluation metrics.