Abstract:Building an accurate load forecasting model with minimal underpredictions is vital to prevent any undesired power outages due to underproduction of electricity. However, the power consumption patterns of the residential sector contain fluctuations and anomalies making them challenging to predict. In this paper, we propose multiple Long Short-Term Memory (LSTM) frameworks with different asymmetric loss functions to impose a higher penalty on underpredictions. We also apply a density-based spatial clustering of applications with noise (DBSCAN) anomaly detection approach, prior to the load forecasting task, to remove any present oultiers. Considering the effect of weather and social factors, seasonality splitting is performed on the three considered datasets from France, Germany, and Hungary containing hourly power consumption, weather, and calendar features. Root-mean-square error (RMSE) results show that removing the anomalies efficiently reduces the underestimation and overestimation errors in all the seasonal datasets. Additionally, asymmetric loss functions and seasonality splitting effectively minimize underestimations despite increasing the overestimation error to some degree. Reducing underpredictions of electricity consumption is essential to prevent power outages that can be damaging to the community.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to improve the accuracy of electricity consumption prediction in residential areas, especially to reduce the situation of underprediction. Specifically, the paper aims to improve the electricity consumption prediction model through the following methods: 1. **Introducing Asymmetric Loss Functions**: Traditional loss functions are usually symmetric, that is, the penalties for overestimation and underestimation are the same. However, in electricity consumption prediction, underestimation may lead to insufficient power supply, which may further cause serious problems such as power outages. Therefore, the paper proposes two different asymmetric loss functions (AL1 and AL2) to impose a higher penalty on underestimation, thereby reducing the underestimation error. 2. **Anomaly Detection and Processing**: There may be outliers in the electricity consumption data, and these outliers will significantly affect the performance of the prediction model. For this reason, the paper uses the density - based clustering algorithm DBSCAN to detect and replace these outliers to improve the accuracy of prediction. 3. **Seasonal Segmentation**: The electricity consumption pattern is affected by weather and social factors and has obvious seasonal fluctuations. The paper performs seasonal segmentation on the data and trains the LSTM model respectively to better capture the electricity consumption characteristics in different seasons, thereby improving the prediction accuracy. ### Specific Problem Description - **Avoiding Insufficient Power Supply**: Underestimating electricity consumption may lead to insufficient power supply, which may further cause problems such as power outages, and this may have a serious impact on the economic and industrial development of the community. - **Reducing Prediction Error**: When dealing with electricity consumption prediction, the traditional symmetric loss function fails to effectively distinguish the effects of overestimation and underestimation, resulting in a large prediction error. - **Handling Abnormal Data**: There may be outliers in the electricity consumption data caused by measurement errors or special events, and these outliers will affect the prediction performance of the model. Through the above methods, the paper aims to construct a more accurate and reliable electricity consumption prediction model to ensure the stable operation of the power system and prevent various adverse effects caused by insufficient power supply.

An Asymmetric Loss with Anomaly Detection LSTM Framework for Power Consumption Prediction

Probabilistic Individual Load Forecasting Using Pinball Loss Guided LSTM

Power Consumption Predicting and Anomaly Detection Based on Long Short-Term Memory Neural Network

A Power Load Forecasting Method Based on Intelligent Data Analysis

A deep learning approach for anomaly detection and prediction in power consumption data

Predicting Energy Consumption Using LSTM, Multi-Layer GRU and Drop-GRU Neural Networks

Anomaly Detection in Electricity Consumption Data using Deep Learning

Electricity Load and Price Forecasting Using a Hybrid Method Based Bidirectional Long Short-Term Memory with Attention Mechanism Model

Comparative study of long short-term memory (LSTM), bidirectional LSTM, and traditional machine learning approaches for energy consumption prediction

Deep Neural Network and Long Short-Term Memory for Electric Power Load Forecasting

Short-term Prediction of Household Electricity Consumption Using Customized LSTM and GRU Models

EMD-Att-LSTM: A Data-driven Strategy Combined with Deep Learning for Short-term Load Forecasting

Analysis and Prediction of Power Distribution Network Loss Based on Machine Learning

Electrical Energy Prediction in Residential Buildings for Short-Term Horizons Using Hybrid Deep Learning Strategy

Efficient Short-Term Electricity Load Forecasting for Effective Energy Management

Short-term Load Forecasting for Distribution Substations Based on Residual Neutral Networks and Long Short-Term Memory Neutral Networks with Attention Mechanism

Power Consumption Predicting and Anomaly Detection Based on Transformer and K-Means

Short term residential load forecasting using long short-term memory recurrent neural network

Demand Forecasting in Smart Grid Using Long Short-Term Memory

A Single Scalable LSTM Model for Short-Term Forecasting of Disaggregated Electricity Loads

Short-term Load Forecasting with Distributed Long Short-Term Memory