{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/multi-horizon-short-term-load-forecasting","title":"Multi-horizon short-term load forecasting using hybrid of LSTM and modified split convolution","arxiv_id":null,"date":"2023-08-15","proceeding":"PeerJ Computer Science 2023 8","authors":["Irshad Ullah","Syed Muhammad Hasanat","Khursheed Aurangzeb","Musaed Alhussein","Muhammad Rizwan","Muhammad Shahid Anwar"],"abstract":"Precise short-term load forecasting (STLF) plays a crucial role in the smooth\r\n operation of power systems, future capacity planning, unit commitment, and\r\n demand response. However, due to its non-stationary and its dependency on\r\nmultiple cyclic and non-cyclic calendric features and non-linear highly correlated\r\n metrological features, an accurate load forecasting with already existing techniques is\r\n challenging. To overcome this challenge, a novel hybrid technique based on long\r\nshort-term memory (LSTM) and a modified split-convolution (SC) neural network\r\n (LSTM-SC) is proposed for single-step and multi-step STLF. The concatenating\r\norder of LSTM and SC in the proposed hybrid network provides an excellent\r\n capability of extraction of sequence-dependent features and other hierarchical spatial\r\nfeatures. The model is evaluated by the Pakistan National Grid load dataset recorded\r\n by the National Transmission and Dispatch Company (NTDC). The load data is preprocessed and multiple other correlated features are incorporated into the data for\r\nperformance enhancement. For generalization capability, the performance of LSTM-SC is evaluated on publicly available datasets of American Electric Power (AEP) and\r\n Independent System Operator New England (ISO-NE). The effect of temperature, a\r\n highly correlated input feature, on load forecasting is investigated either by removing\r\nthe temperature or adding a Gaussian random noise into it. The performance\r\n evaluation in terms of RMSE, MAE, and MAPE of the proposed model on the NTDC\r\ndataset are 500.98, 372.62, and 3.72% for multi-step while 322.90, 244.22, and 2.38%\r\n for single-step load forecasting. The result shows that the proposed method has less\r\nforecasting error, strong generalization capability, and satisfactory performance on\r\nmulti-horizon.","url_abs":"https://peerj.com/articles/cs-1487/","url_pdf":"https://doi.org/10.7717/peerj-cs.1487","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"multi-horizon-short-term-load-forecasting","repo_url":"https://github.com/SyedHasnat/Papers","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"data-ablation","task_name":"Data Ablation"},{"task_slug":"load-forecasting","task_name":"Load Forecasting"},{"task_slug":"missing-elements","task_name":"Missing Elements"},{"task_slug":"multivariate-time-series-forecasting","task_name":"Multivariate Time Series Forecasting"},{"task_slug":"outlier-interpretation","task_name":"Outlier Interpretation"},{"task_slug":"time-series-forecasting","task_name":"Time Series Forecasting"},{"task_slug":"univariate-time-series-forecasting","task_name":"Univariate Time Series Forecasting"}],"methods":[{"method_slug":"1d-cnn","method_name":"1D CNN"},{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":null,"method_name":"American"},{"method_slug":"auxiliary-classifier","method_name":"Auxiliary Classifier"},{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"electric","method_name":"Electric"},{"method_slug":"googlenet","method_name":"GoogLeNet"},{"method_slug":"inception-module","method_name":"Inception Module"},{"method_slug":"lstm","method_name":"LSTM"},{"method_slug":"local-response-normalization","method_name":"Local Response Normalization"},{"method_slug":"mae","method_name":"MAE"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"tanh-activation","method_name":"Tanh Activation"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/multivariate-time-series-forecasting-on-aep","task":"Multivariate Time Series Forecasting","dataset":"AEP","model":"LSTM-SC","rank_in_archive_order":1,"of":1,"metrics":{"12 steps MAPE":"2.58","12 steps RMSE":"549.92"},"uses_additional_data":false},{"leaderboard":"/sota/univariate-time-series-forecasting-on-aep","task":"Univariate Time Series Forecasting","dataset":"AEP","model":"LSTM-SC","rank_in_archive_order":1,"of":1,"metrics":{"MAPE (%)":"0.67","RMSE":"97.78"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}