Icon

Energy Usage Prediction with LSTM (Univariate TSA)

Electric Usage Prediction Using LSTM

There has been no description set for this workflow's metadata.

UNIVARIATE TIME SERIES ANALYSISElectric Usage Prediction Using LSTMThis workflow implement Long Short Term Memory (LSTM) model topredict electric consumptionDataset: https://www.kaggle.com/datasets/jaganadhg/house-hold-energy-data Data Loading & preprocessing Model Training & Prediction Model Evaluation Performance Workflow summary:1. Data loading2. Data cleaning 3. Timestemp alignment verified component used to align data by hour4. Fill in missing value5. Partitioning - Split data to training and validation dataset by 80:20 ratio6. Input vector for keras preparation for both training and testing data7. Partitioning - Split validation dataset to validation and test data by 50:50 in ratio8. Model architecture 9. Model Training using keras network learner10. Execute prediction using Keras network executor11. Append predicted value and test data12. Evaluate model performance using Numeric Scorer13. Plot Actual vs Predicted value Read dataRemoveunnecessarycolumnSplit datasetto train & test(80:20)training withMSE as loss functionInput Shape[timestep, feature][7, 1]LSTM LayerInput Shape[7,1]Activation Tanh[1] tensor foroutputActivation ReLUvisualise thedatavalidation:test(50:50)execute the modelNode 535format date(str)align timestep by hourfilter date columnfill in missing valueappend test data& predicted valueModel Quality MetricJoin with "date" columnNormalize train datanormalize validation datadenormalize test dataLSTM LayerActivation TanhNode 555 Excel Reader Column Filter Partitioning RestructureTraining Set Keras NetworkLearner Keras Input Layer Keras LSTM Layer Keras Dense Layer Line Plot RestructureTraining Set Partitioning Keras NetworkExecutor Column Aggregator String to Date&Time Timestamp Alignment Column Filter Missing Value Column Appender Numeric Scorer Joiner Normalizer Normalizer (Apply) Metanode Denormalizer Keras LSTM Layer Box Plot Keras NetworkWriter UNIVARIATE TIME SERIES ANALYSISElectric Usage Prediction Using LSTMThis workflow implement Long Short Term Memory (LSTM) model topredict electric consumptionDataset: https://www.kaggle.com/datasets/jaganadhg/house-hold-energy-data Data Loading & preprocessing Model Training & Prediction Model Evaluation Performance Workflow summary:1. Data loading2. Data cleaning 3. Timestemp alignment verified component used to align data by hour4. Fill in missing value5. Partitioning - Split data to training and validation dataset by 80:20 ratio6. Input vector for keras preparation for both training and testing data7. Partitioning - Split validation dataset to validation and test data by 50:50 in ratio8. Model architecture 9. Model Training using keras network learner10. Execute prediction using Keras network executor11. Append predicted value and test data12. Evaluate model performance using Numeric Scorer13. Plot Actual vs Predicted value Read dataRemoveunnecessarycolumnSplit datasetto train & test(80:20)training withMSE as loss functionInput Shape[timestep, feature][7, 1]LSTM LayerInput Shape[7,1]Activation Tanh[1] tensor foroutputActivation ReLUvisualise thedatavalidation:test(50:50)execute the modelNode 535format date(str)align timestep by hourfilter date columnfill in missing valueappend test data& predicted valueModel Quality MetricJoin with "date" columnNormalize train datanormalize validation datadenormalize test dataLSTM LayerActivation TanhNode 555 Excel Reader Column Filter Partitioning RestructureTraining Set Keras NetworkLearner Keras Input Layer Keras LSTM Layer Keras Dense Layer Line Plot RestructureTraining Set Partitioning Keras NetworkExecutor Column Aggregator String to Date&Time Timestamp Alignment Column Filter Missing Value Column Appender Numeric Scorer Joiner Normalizer Normalizer (Apply) Metanode Denormalizer Keras LSTM Layer Box Plot Keras NetworkWriter

Nodes

Extensions

Links

Download

To use this workflow in KNIME, download it from the below URL and open it in KNIME:

Download Workflow
Created by: USER
Created at: 2022-06-10
On NodePit since: 2023-04-07
Last update: 2024-11-23
Created with KNIME version: v4.6.1

Deploy, schedule, execute, and monitor your KNIME workflows locally, in the cloud or on-premises – with our brand new NodePit Runner.

Try NodePit Runner!