Skip to content

Implementation of a neural network algorithm for forecasting water consumption in homes.

Notifications You must be signed in to change notification settings

gchacaltana/forecasting_neural_networks

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Pronóstico con redes neuronales

Resumen

Aplicación demo que implementa el proceso de entrenamiento de una red neuronal para un caso de uso de pronóstico de consumo mensual de agua potable de una vivienda.

Conjunto de datos

El conjunto de datos corresponde al consumo de agua potable (en m3) de una vivienda. Las variables son: Fecha Facturación, Año, Mes, Consumo M3 e Importe Total.

Ejemplo de estructura de información.

2017-11-11         2017   11  27.957   S/. 42.63
2017-12-12         2017   12  27.961   S/. 32.09
2018-01-11         2018    1  31.906   S/. 41.17
2018-02-10         2018    2  38.824   S/. 47.41
2018-03-13         2018    3  38.417   S/. 60.94
2018-04-13         2018    4  33.981   S/. 67.16
2018-05-12         2018    5  39.217   S/. 78.19
2018-06-12         2018    6  40.855   S/. 76.69
2018-07-13         2018    7  38.783   S/. 72.53
2018-08-11         2018    8  32.925   S/. 68.48
2018-09-11         2018    9  37.412   S/. 80.35
2018-10-11         2018   10  42.413   S/. 77.30
2018-11-10         2018   11  45.599   S/. 85.96
2018-12-12         2018   12  47.315  S/. 122.53
2019-01-10         2019    1  28.595   S/. 75.69
2019-02-09         2019    2  38.238  S/. 105.69
2019-03-12         2019    3  46.413  S/. 130.57
2019-04-11         2019    4  37.242  S/. 104.81
2019-05-11         2019    5  29.696   S/. 83.58
2019-06-12         2019    6  36.779  S/. 103.57
2019-07-11         2019    7  32.506   S/. 79.34
2019-08-12         2019    8  32.848   S/. 88.09
2019-09-10         2019    9  35.495  S/. 100.91
2019-10-11         2019   10  38.442  S/. 109.31
2019-11-11         2019   11  40.088  S/. 115.42
2019-12-11         2019   12  35.821  S/. 103.13
2020-01-10         2020    1  35.784  S/. 103.02
2020-02-10         2020    2  35.368  S/. 101.80
2020-03-11         2020    3  41.188  S/. 118.52
2020-04-11         2020    4  34.664   S/. 92.39
2020-05-12         2020    5  32.021   S/. 90.84
2020-06-10         2020    6  30.789   S/. 88.68
2020-07-10         2020    7  29.120   S/. 87.96
2020-08-10         2020    8  29.894   S/. 91.42
2020-09-10         2020    9  27.399   S/. 78.88
2020-10-10         2020   10  25.238   S/. 72.69
2020-11-11         2020   11  22.726   S/. 65.43

Configuración Red Neuronal

Creamos modelo de red neuronal feed forward.
Arquitectura:
    * 01 capa oculta con "n" neuronas (ingresadas por consola)
    * 01 neurona de salida.
Función de activación: Tangente Hiperbólica.(para valores normalizados -1 a 1)
Optimizador: Adam
Métrica de Pérdida: (Loss) Error Absoluto Medio
Para calcular el acuracy, se utilizará  Error Cuadrático Medio (MSE)

Requisitos

* numpy 1.19.3
* pandas 1.0.1   
* Keras 2.4.3
* tensorflow 2.4.0
* scikit-learn 0.22.1
* matplotlib 3.1.3

Ejecución

python Application.py train

About

Implementation of a neural network algorithm for forecasting water consumption in homes.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Languages