0 / 0
Torna alla versione inglese della documentazione
Utilizzare i modelli di fondazione delle serie temporali IBM Granite e l'API di previsione per prevedere i valori futuri (beta)
Ultimo aggiornamento: 18 dic 2024
Utilizzare i modelli di fondazione delle serie temporali IBM Granite e l'API di previsione per prevedere i valori futuri (beta)

Utilizzate le API di previsione delle serie temporali e i modelli di serie temporali IBM Granite disponibili su watsonx.ai per prevedere i valori futuri sulla base dei dati storici.

Importante: il metodo di previsione delle serie temporali dell'API watsonx.ai è disponibile come funzione beta.

Riferimento API

Per maggiori dettagli, consultare la documentazione di riferimento dell'API.

SDK Python

È possibile scrivere codice che utilizza i modelli di serie temporali per prevedere i dati utilizzando la libreria Python watsonx.ai Per ulteriori informazioni, consultare i seguenti riferimenti:

Notebook di esempio

Utilizzare i modelli di fondazione delle serie temporali e i dati delle serie temporali per prevedere la domanda di energia

Previsione delle serie temporali

Utilizzare l'API delle serie temporali per passare le osservazioni dei dati storici a un foundation model delle serie temporali in grado di prevedere i valori futuri con un'inferenza a zero colpi. Ad esempio, è possibile utilizzare questo metodo per prevedere i valori futuri in base alle osservazioni storiche dei seguenti tipi di dati:

  • Prezzi delle azioni e volumi di scambio
  • Dati di elettrocardiogramma (ECG) o registrazioni di polisonnogramma (PSG)
  • Dati di temperatura o sismografici
  • Misurazioni delle prestazioni della rete

Modelli di fondazione supportati

I modelli IBM Granite time series foundation con i seguenti ID modello sono disponibili per essere utilizzati per le previsioni in watsonx.ai:

  • ibm/granite-ttm-512-96-r2: Richiede almeno 512 punti dati per ogni set di dati.
  • ibm/granite-ttm-1024-96-r2: Richiede almeno 1.024 punti dati per ogni set di dati.
  • ibm/granite-ttm-1536-96-r2: Richiede almeno 1.536 punti dati per ogni set di dati.

I modelli di serie temporali Granite funzionano al meglio con punti dati in intervalli di minuti o ore e generano un set di dati di previsione con 96 punti dati per colonna target. Per ottenere risultati ottimali, utilizzate il modello che prende il maggior numero di punti dati in base ai dati a vostra disposizione.

Per ulteriori informazioni sui modelli, vedere Modelli di fondazioneIBM.

Per ottenere programmaticamente un elenco dei modelli di serie temporali disponibili per l'uso, è possibile utilizzare il metodo List the available foundation models nell'API watsonx.ai as a service. Specificare il parametro 'filters=function_time_series_forecast per restituire solo i modelli di serie temporali disponibili.

Ad esempio:

curl -X GET \
  '{url}/ml/v1/foundation_model_specs?version=2024-11-14&filters=function_time_series_forecast'

Nell'esempio, sostituire la variabile '{url} con il valore giusto per la propria istanza, ad esempio 'us-south.ml.cloud.ibm.com.

Requisiti di dati

Una serie temporale è un insieme di punti di dati raccolti nel tempo.

Per utilizzare i modelli di serie temporali Granite per prevedere i valori futuri, i dati inviati per l'analisi devono soddisfare i seguenti requisiti:

  • I dati registrati devono essere numerici, come le temperature o i prezzi delle azioni.

  • I dati devono includere un numero sufficiente di punti di contesto storico affinché il modello sia in grado di fare una previsione. Ogni foundation model delle serie temporali specifica un numero minimo di punti dati per ogni set di dati. Se si specifica un numero superiore a quello richiesto, il modello utilizza i punti dati più recenti fino al requisito del modello e ignora gli altri.

  • Ogni set di dati deve avere un numero uguale di elementi negli array specificati per i " date, " id_columns e " target_columns inclusi nel corpo della richiesta. Non è possibile saltare un punto dati o specificare " null come punto dati. Controllare i dati prima di inviare la richiesta.

  • Sebbene il servizio accetti i formati di data e ora più comuni, se si specificano date e orari nel formato ISO 8601 (2024-11-12T15:06:35), si evita la confusione che può sorgere a causa delle differenze nelle convenzioni di formattazione delle date. Ad esempio, 11/12/2024 significa 12 novembre o 11 dicembre? Specificare le date nel formato ISO 8601 con l'offset del tempo universale coordinato (UTC) (2024-11-12T15:06:35+0000) per evitare duplicazioni o mancanza di timestamp nei risultati di previsione generati.

  • Campionare i dati con una frequenza uniforme. Ad esempio, i dati possono essere osservati con incrementi di 1 minuto, 1 ora o 1 giorno. Se i timestamp non sono uniformi non viene generato alcun errore, ma la qualità dei risultati potrebbe essere scarsa. I dati di previsione generati sono anche formattati con la frequenza specificata con il parametro " freq.

    Ad esempio, se si specifica una frequenza di un giorno ("freq":"1D"), le marche temporali avranno probabilmente un aspetto simile a questo:

    "date": [
          "2024-11-15T15:06:35",
          "2024-11-16T15:06:35",
    ...
    ]
    

    Se la frequenza è di 5 minuti ("freq":"5min"), ogni data e ora saranno probabilmente distanti 5 minuti l'una dall'altra.

    "date": [
          "2024-11-15T15:06:35",
          "2024-11-15T15:11:35",
    ...
    ]
    

    Per i valori supportati per le abbreviazioni di data e ora nel parametro frequenza, vedere Alias di periodo nella documentazione della libreria pandas.

  • Se si includono punti di dati provenienti da più fonti per l'analisi, ogni punto di dati deve fornire un identificatore univoco per indicare a quale set di dati appartiene.

    Ad esempio, se state analizzando le tendenze dello shopping, potreste essere interessati ai dati di vendita di più negozi durante le festività natalizie. La tabella seguente mostra due set di dati, A e B. Il dataset A mostra le vendite giornaliere per la località A e il dataset B mostra le vendite giornaliere per la località B.

    Tabella 1. Esempio di dati di vendita
    Data Dataset Totale vendite (USD)
    7 dicembre 2023 A $24,988
    7 dicembre 2023 B $63,788
    8 dicembre 2023 A $41,855
    8 dicembre 2023 B $105,678
    ... ... ...

    Anche se la tabella mostra le osservazioni di due giorni, la richiesta deve includere da 512 a 1.536 punti dati, a seconda del modello utilizzato. Quando i dati della tabella vengono inviati all'API, sono formattati come segue:

    ...
    "data": {
    "date": [
      "2024-12-07T00:00:00",
      "2024-12-07T00:00:00",
      "2024-12-08T00:00:00",
      "2024-12-08T00:00:00",
      ...
    ],
    "ids": [
      "A",
      "B",
      "A",
      "B",
      ...
    ],
    "sales-usd": [
      24988,
      63788,
      41855,
      105678,
      ...
    ]
    }
    
  • È possibile fare previsioni utilizzando punti di dati multivariati, ossia punti di dati che misurano fattori diversi durante lo stesso intervallo di tempo.

    Ad esempio, potreste essere interessati sia alle vendite che ai resi di un negozio durante le festività natalizie. La tabella seguente mostra alcune voci di due set di dati. Un set di dati mostra le vendite al giorno e l'altro set di dati mostra i ritorni al giorno.

    Tabella 2. Esempio di dati sulle vendite e sui resi
    Data Vendite (USD) Rendimenti (USD)
    7 dicembre 2023 $63,788 $14,788
    8 dicembre 2023 $105,678 $25,678
    ... ... ...

    Anche se la tabella mostra le osservazioni di due giorni, la richiesta deve includere da 512 a 1.536 punti dati, a seconda del modello utilizzato. Quando i dati della tabella vengono inviati all'API, sono formattati come segue:

    ...
    "data": {
    "date": [
      "2024-12-07T00:00:00",
      "2024-12-07T00:00:00",
      "2024-12-08T00:00:00",
      "2024-12-08T00:00:00",
      ...
    ],
    "ids": [
      "sales",
      "returns",
      "sales",
      "returns",
      ...
    ],
    "sales": [
      63788,
      105678,
      ...
    ],
    "returns": [
      14788,
      25678
      ...
    ]
    }
    

Esempio

Questo esempio utilizza il modello " granite-ttm-512-96-r2 per prevedere il consumo di energia in kilowattora all'ora sulla base di dati orari per il periodo compreso tra il 1° agosto e il 22 agosto 2024.

Vengono inviati i dati di un solo set di dati, il che significa che l'oggetto opzionale " id_columns può essere omesso dallo schema. Se viene utilizzato più di un set di dati, l'oggetto " id_columns deve essere incluso.

Il corpo della richiesta è diviso in due sezioni:

  • Schema: Specifica quali set di dati saranno inclusi e quali campi analizzare.
  • Dati: I dati da analizzare.

Lo schema si presenta come segue:

"schema": {
    "timestamp_column": "date",
    "freq":"1h",
    "target_columns": [
      "consumption-kwh"
    ]
  },

I dati comprendono le informazioni riportate nella tabella seguente.

Tabella 3. Esempio di consumo energetico
Data e ora (data) Consumo orario di energia in kWh (consumo-kwh)
2024-08-01T00:00:00 1.343
2024-08-01T01:00:00 1.274
2024-08-01T02:00:00 1.126

Il modello richiede almeno 512 punti di dati per ogni set di dati per garantire che ci siano dati sufficienti per il modello per prevedere i valori futuri. Sebbene la tabella mostri 3 punti di dati, il corpo della richiesta contiene 528 punti di dati.

Nell'esempio seguente, sostituire{url}variable with the right value for your instance, such asus-south.ml.cloud.ibm.com. Aggiungere il proprio token portatore e l'ID del progetto.

Esempio di richiesta REST

curl -X POST \
  'https://{region}.ml.cloud.ibm.com/ml/v1/time_series/forecast?version=2024-11-15' \
  --header 'Accept: application/json' \
  --header 'Content-Type: application/json' \
  --header 'Authorization: Bearer eyJraWQiOi...' \
  --data '{
      "model_id":"
    }'

Il payload dei dati contiene il seguente frammento JSON:

{
  "model_id":"ibm/granite-ttm-512-96-r2",
  "project_id":"4ab1437e-4c9d-47cf-9926-9ab52db1ee01",
  "schema": {
    "timestamp_column": "date",
            "freq":"1h",
            "target_columns": [
                "consumption-kwh"
            ]
        },
  "data": {
        "date": [
            "2024-08-01T00:00:00",
"2024-08-01T01:00:00",
"2024-08-01T02:00:00",
"2024-08-01T03:00:00",
"2024-08-01T04:00:00",
"2024-08-01T05:00:00",
"2024-08-01T06:00:00",
"2024-08-01T07:00:00",
"2024-08-01T08:00:00",
"2024-08-01T09:00:00",
"2024-08-01T10:00:00",
"2024-08-01T11:00:00",
"2024-08-01T12:00:00",
"2024-08-01T13:00:00",
"2024-08-01T14:00:00",
"2024-08-01T15:00:00",
"2024-08-01T16:00:00",
"2024-08-01T17:00:00",
"2024-08-01T18:00:00",
"2024-08-01T19:00:00",
"2024-08-01T20:00:00",
"2024-08-01T21:00:00",
"2024-08-01T22:00:00",
"2024-08-01T23:00:00",
"2024-08-02T00:00:00",
"2024-08-02T01:00:00",
"2024-08-02T02:00:00",
"2024-08-02T03:00:00",
"2024-08-02T04:00:00",
"2024-08-02T05:00:00",
"2024-08-02T06:00:00",
"2024-08-02T07:00:00",
"2024-08-02T08:00:00",
"2024-08-02T09:00:00",
"2024-08-02T10:00:00",
"2024-08-02T11:00:00",
"2024-08-02T12:00:00",
"2024-08-02T13:00:00",
"2024-08-02T14:00:00",
"2024-08-02T15:00:00",
"2024-08-02T16:00:00",
"2024-08-02T17:00:00",
"2024-08-02T18:00:00",
"2024-08-02T19:00:00",
"2024-08-02T20:00:00",
"2024-08-02T21:00:00",
"2024-08-02T22:00:00",
"2024-08-02T23:00:00",
"2024-08-03T00:00:00",
"2024-08-03T01:00:00",
"2024-08-03T02:00:00",
"2024-08-03T03:00:00",
"2024-08-03T04:00:00",
"2024-08-03T05:00:00",
"2024-08-03T06:00:00",
"2024-08-03T07:00:00",
"2024-08-03T08:00:00",
"2024-08-03T09:00:00",
"2024-08-03T10:00:00",
"2024-08-03T11:00:00",
"2024-08-03T12:00:00",
"2024-08-03T13:00:00",
"2024-08-03T14:00:00",
"2024-08-03T15:00:00",
"2024-08-03T16:00:00",
"2024-08-03T17:00:00",
"2024-08-03T18:00:00",
"2024-08-03T19:00:00",
"2024-08-03T20:00:00",
"2024-08-03T21:00:00",
"2024-08-03T22:00:00",
"2024-08-03T23:00:00",
"2024-08-04T00:00:00",
"2024-08-04T01:00:00",
"2024-08-04T02:00:00",
"2024-08-04T03:00:00",
"2024-08-04T04:00:00",
"2024-08-04T05:00:00",
"2024-08-04T06:00:00",
"2024-08-04T07:00:00",
"2024-08-04T08:00:00",
"2024-08-04T09:00:00",
"2024-08-04T10:00:00",
"2024-08-04T11:00:00",
"2024-08-04T12:00:00",
"2024-08-04T13:00:00",
"2024-08-04T14:00:00",
"2024-08-04T15:00:00",
"2024-08-04T16:00:00",
"2024-08-04T17:00:00",
"2024-08-04T18:00:00",
"2024-08-04T19:00:00",
"2024-08-04T20:00:00",
"2024-08-04T21:00:00",
"2024-08-04T22:00:00",
"2024-08-04T23:00:00",
"2024-08-05T00:00:00",
"2024-08-05T01:00:00",
"2024-08-05T02:00:00",
"2024-08-05T03:00:00",
"2024-08-05T04:00:00",
"2024-08-05T05:00:00",
"2024-08-05T06:00:00",
"2024-08-05T07:00:00",
"2024-08-05T08:00:00",
"2024-08-05T09:00:00",
"2024-08-05T10:00:00",
"2024-08-05T11:00:00",
"2024-08-05T12:00:00",
"2024-08-05T13:00:00",
"2024-08-05T14:00:00",
"2024-08-05T15:00:00",
"2024-08-05T16:00:00",
"2024-08-05T17:00:00",
"2024-08-05T18:00:00",
"2024-08-05T19:00:00",
"2024-08-05T20:00:00",
"2024-08-05T21:00:00",
"2024-08-05T22:00:00",
"2024-08-05T23:00:00",
"2024-08-06T00:00:00",
"2024-08-06T01:00:00",
"2024-08-06T02:00:00",
"2024-08-06T03:00:00",
"2024-08-06T04:00:00",
"2024-08-06T05:00:00",
"2024-08-06T06:00:00",
"2024-08-06T07:00:00",
"2024-08-06T08:00:00",
"2024-08-06T09:00:00",
"2024-08-06T10:00:00",
"2024-08-06T11:00:00",
"2024-08-06T12:00:00",
"2024-08-06T13:00:00",
"2024-08-06T14:00:00",
"2024-08-06T15:00:00",
"2024-08-06T16:00:00",
"2024-08-06T17:00:00",
"2024-08-06T18:00:00",
"2024-08-06T19:00:00",
"2024-08-06T20:00:00",
"2024-08-06T21:00:00",
"2024-08-06T22:00:00",
"2024-08-06T23:00:00",
"2024-08-07T00:00:00",
"2024-08-07T01:00:00",
"2024-08-07T02:00:00",
"2024-08-07T03:00:00",
"2024-08-07T04:00:00",
"2024-08-07T05:00:00",
"2024-08-07T06:00:00",
"2024-08-07T07:00:00",
"2024-08-07T08:00:00",
"2024-08-07T09:00:00",
"2024-08-07T10:00:00",
"2024-08-07T11:00:00",
"2024-08-07T12:00:00",
"2024-08-07T13:00:00",
"2024-08-07T14:00:00",
"2024-08-07T15:00:00",
"2024-08-07T16:00:00",
"2024-08-07T17:00:00",
"2024-08-07T18:00:00",
"2024-08-07T19:00:00",
"2024-08-07T20:00:00",
"2024-08-07T21:00:00",
"2024-08-07T22:00:00",
"2024-08-07T23:00:00",
"2024-08-08T00:00:00",
"2024-08-08T01:00:00",
"2024-08-08T02:00:00",
"2024-08-08T03:00:00",
"2024-08-08T04:00:00",
"2024-08-08T05:00:00",
"2024-08-08T06:00:00",
"2024-08-08T07:00:00",
"2024-08-08T08:00:00",
"2024-08-08T09:00:00",
"2024-08-08T10:00:00",
"2024-08-08T11:00:00",
"2024-08-08T12:00:00",
"2024-08-08T13:00:00",
"2024-08-08T14:00:00",
"2024-08-08T15:00:00",
"2024-08-08T16:00:00",
"2024-08-08T17:00:00",
"2024-08-08T18:00:00",
"2024-08-08T19:00:00",
"2024-08-08T20:00:00",
"2024-08-08T21:00:00",
"2024-08-08T22:00:00",
"2024-08-08T23:00:00",
"2024-08-09T00:00:00",
"2024-08-09T01:00:00",
"2024-08-09T02:00:00",
"2024-08-09T03:00:00",
"2024-08-09T04:00:00",
"2024-08-09T05:00:00",
"2024-08-09T06:00:00",
"2024-08-09T07:00:00",
"2024-08-09T08:00:00",
"2024-08-09T09:00:00",
"2024-08-09T10:00:00",
"2024-08-09T11:00:00",
"2024-08-09T12:00:00",
"2024-08-09T13:00:00",
"2024-08-09T14:00:00",
"2024-08-09T15:00:00",
"2024-08-09T16:00:00",
"2024-08-09T17:00:00",
"2024-08-09T18:00:00",
"2024-08-09T19:00:00",
"2024-08-09T20:00:00",
"2024-08-09T21:00:00",
"2024-08-09T22:00:00",
"2024-08-09T23:00:00",
"2024-08-10T00:00:00",
"2024-08-10T01:00:00",
"2024-08-10T02:00:00",
"2024-08-10T03:00:00",
"2024-08-10T04:00:00",
"2024-08-10T05:00:00",
"2024-08-10T06:00:00",
"2024-08-10T07:00:00",
"2024-08-10T08:00:00",
"2024-08-10T09:00:00",
"2024-08-10T10:00:00",
"2024-08-10T11:00:00",
"2024-08-10T12:00:00",
"2024-08-10T13:00:00",
"2024-08-10T14:00:00",
"2024-08-10T15:00:00",
"2024-08-10T16:00:00",
"2024-08-10T17:00:00",
"2024-08-10T18:00:00",
"2024-08-10T19:00:00",
"2024-08-10T20:00:00",
"2024-08-10T21:00:00",
"2024-08-10T22:00:00",
"2024-08-10T23:00:00",
"2024-08-11T00:00:00",
"2024-08-11T01:00:00",
"2024-08-11T02:00:00",
"2024-08-11T03:00:00",
"2024-08-11T04:00:00",
"2024-08-11T05:00:00",
"2024-08-11T06:00:00",
"2024-08-11T07:00:00",
"2024-08-11T08:00:00",
"2024-08-11T09:00:00",
"2024-08-11T10:00:00",
"2024-08-11T11:00:00",
"2024-08-11T12:00:00",
"2024-08-11T13:00:00",
"2024-08-11T14:00:00",
"2024-08-11T15:00:00",
"2024-08-11T16:00:00",
"2024-08-11T17:00:00",
"2024-08-11T18:00:00",
"2024-08-11T19:00:00",
"2024-08-11T20:00:00",
"2024-08-11T21:00:00",
"2024-08-11T22:00:00",
"2024-08-11T23:00:00",
"2024-08-12T00:00:00",
"2024-08-12T01:00:00",
"2024-08-12T02:00:00",
"2024-08-12T03:00:00",
"2024-08-12T04:00:00",
"2024-08-12T05:00:00",
"2024-08-12T06:00:00",
"2024-08-12T07:00:00",
"2024-08-12T08:00:00",
"2024-08-12T09:00:00",
"2024-08-12T10:00:00",
"2024-08-12T11:00:00",
"2024-08-12T12:00:00",
"2024-08-12T13:00:00",
"2024-08-12T14:00:00",
"2024-08-12T15:00:00",
"2024-08-12T16:00:00",
"2024-08-12T17:00:00",
"2024-08-12T18:00:00",
"2024-08-12T19:00:00",
"2024-08-12T20:00:00",
"2024-08-12T21:00:00",
"2024-08-12T22:00:00",
"2024-08-12T23:00:00",
"2024-08-13T00:00:00",
"2024-08-13T01:00:00",
"2024-08-13T02:00:00",
"2024-08-13T03:00:00",
"2024-08-13T04:00:00",
"2024-08-13T05:00:00",
"2024-08-13T06:00:00",
"2024-08-13T07:00:00",
"2024-08-13T08:00:00",
"2024-08-13T09:00:00",
"2024-08-13T10:00:00",
"2024-08-13T11:00:00",
"2024-08-13T12:00:00",
"2024-08-13T13:00:00",
"2024-08-13T14:00:00",
"2024-08-13T15:00:00",
"2024-08-13T16:00:00",
"2024-08-13T17:00:00",
"2024-08-13T18:00:00",
"2024-08-13T19:00:00",
"2024-08-13T20:00:00",
"2024-08-13T21:00:00",
"2024-08-13T22:00:00",
"2024-08-13T23:00:00",
"2024-08-14T00:00:00",
"2024-08-14T01:00:00",
"2024-08-14T02:00:00",
"2024-08-14T03:00:00",
"2024-08-14T04:00:00",
"2024-08-14T05:00:00",
"2024-08-14T06:00:00",
"2024-08-14T07:00:00",
"2024-08-14T08:00:00",
"2024-08-14T09:00:00",
"2024-08-14T10:00:00",
"2024-08-14T11:00:00",
"2024-08-14T12:00:00",
"2024-08-14T13:00:00",
"2024-08-14T14:00:00",
"2024-08-14T15:00:00",
"2024-08-14T16:00:00",
"2024-08-14T17:00:00",
"2024-08-14T18:00:00",
"2024-08-14T19:00:00",
"2024-08-14T20:00:00",
"2024-08-14T21:00:00",
"2024-08-14T22:00:00",
"2024-08-14T23:00:00",
"2024-08-15T00:00:00",
"2024-08-15T01:00:00",
"2024-08-15T02:00:00",
"2024-08-15T03:00:00",
"2024-08-15T04:00:00",
"2024-08-15T05:00:00",
"2024-08-15T06:00:00",
"2024-08-15T07:00:00",
"2024-08-15T08:00:00",
"2024-08-15T09:00:00",
"2024-08-15T10:00:00",
"2024-08-15T11:00:00",
"2024-08-15T12:00:00",
"2024-08-15T13:00:00",
"2024-08-15T14:00:00",
"2024-08-15T15:00:00",
"2024-08-15T16:00:00",
"2024-08-15T17:00:00",
"2024-08-15T18:00:00",
"2024-08-15T19:00:00",
"2024-08-15T20:00:00",
"2024-08-15T21:00:00",
"2024-08-15T22:00:00",
"2024-08-15T23:00:00",
"2024-08-16T00:00:00",
"2024-08-16T01:00:00",
"2024-08-16T02:00:00",
"2024-08-16T03:00:00",
"2024-08-16T04:00:00",
"2024-08-16T05:00:00",
"2024-08-16T06:00:00",
"2024-08-16T07:00:00",
"2024-08-16T08:00:00",
"2024-08-16T09:00:00",
"2024-08-16T10:00:00",
"2024-08-16T11:00:00",
"2024-08-16T12:00:00",
"2024-08-16T13:00:00",
"2024-08-16T14:00:00",
"2024-08-16T15:00:00",
"2024-08-16T16:00:00",
"2024-08-16T17:00:00",
"2024-08-16T18:00:00",
"2024-08-16T19:00:00",
"2024-08-16T20:00:00",
"2024-08-16T21:00:00",
"2024-08-16T22:00:00",
"2024-08-16T23:00:00",
"2024-08-17T00:00:00",
"2024-08-17T01:00:00",
"2024-08-17T02:00:00",
"2024-08-17T03:00:00",
"2024-08-17T04:00:00",
"2024-08-17T05:00:00",
"2024-08-17T06:00:00",
"2024-08-17T07:00:00",
"2024-08-17T08:00:00",
"2024-08-17T09:00:00",
"2024-08-17T10:00:00",
"2024-08-17T11:00:00",
"2024-08-17T12:00:00",
"2024-08-17T13:00:00",
"2024-08-17T14:00:00",
"2024-08-17T15:00:00",
"2024-08-17T16:00:00",
"2024-08-17T17:00:00",
"2024-08-17T18:00:00",
"2024-08-17T19:00:00",
"2024-08-17T20:00:00",
"2024-08-17T21:00:00",
"2024-08-17T22:00:00",
"2024-08-17T23:00:00",
"2024-08-18T00:00:00",
"2024-08-18T01:00:00",
"2024-08-18T02:00:00",
"2024-08-18T03:00:00",
"2024-08-18T04:00:00",
"2024-08-18T05:00:00",
"2024-08-18T06:00:00",
"2024-08-18T07:00:00",
"2024-08-18T08:00:00",
"2024-08-18T09:00:00",
"2024-08-18T10:00:00",
"2024-08-18T11:00:00",
"2024-08-18T12:00:00",
"2024-08-18T13:00:00",
"2024-08-18T14:00:00",
"2024-08-18T15:00:00",
"2024-08-18T16:00:00",
"2024-08-18T17:00:00",
"2024-08-18T18:00:00",
"2024-08-18T19:00:00",
"2024-08-18T20:00:00",
"2024-08-18T21:00:00",
"2024-08-18T22:00:00",
"2024-08-18T23:00:00",
"2024-08-19T00:00:00",
"2024-08-19T01:00:00",
"2024-08-19T02:00:00",
"2024-08-19T03:00:00",
"2024-08-19T04:00:00",
"2024-08-19T05:00:00",
"2024-08-19T06:00:00",
"2024-08-19T07:00:00",
"2024-08-19T08:00:00",
"2024-08-19T09:00:00",
"2024-08-19T10:00:00",
"2024-08-19T11:00:00",
"2024-08-19T12:00:00",
"2024-08-19T13:00:00",
"2024-08-19T14:00:00",
"2024-08-19T15:00:00",
"2024-08-19T16:00:00",
"2024-08-19T17:00:00",
"2024-08-19T18:00:00",
"2024-08-19T19:00:00",
"2024-08-19T20:00:00",
"2024-08-19T21:00:00",
"2024-08-19T22:00:00",
"2024-08-19T23:00:00",
"2024-08-20T00:00:00",
"2024-08-20T01:00:00",
"2024-08-20T02:00:00",
"2024-08-20T03:00:00",
"2024-08-20T04:00:00",
"2024-08-20T05:00:00",
"2024-08-20T06:00:00",
"2024-08-20T07:00:00",
"2024-08-20T08:00:00",
"2024-08-20T09:00:00",
"2024-08-20T10:00:00",
"2024-08-20T11:00:00",
"2024-08-20T12:00:00",
"2024-08-20T13:00:00",
"2024-08-20T14:00:00",
"2024-08-20T15:00:00",
"2024-08-20T16:00:00",
"2024-08-20T17:00:00",
"2024-08-20T18:00:00",
"2024-08-20T19:00:00",
"2024-08-20T20:00:00",
"2024-08-20T21:00:00",
"2024-08-20T22:00:00",
"2024-08-20T23:00:00",
"2024-08-21T00:00:00",
"2024-08-21T01:00:00",
"2024-08-21T02:00:00",
"2024-08-21T03:00:00",
"2024-08-21T04:00:00",
"2024-08-21T05:00:00",
"2024-08-21T06:00:00",
"2024-08-21T07:00:00",
"2024-08-21T08:00:00",
"2024-08-21T09:00:00",
"2024-08-21T10:00:00",
"2024-08-21T11:00:00",
"2024-08-21T12:00:00",
"2024-08-21T13:00:00",
"2024-08-21T14:00:00",
"2024-08-21T15:00:00",
"2024-08-21T16:00:00",
"2024-08-21T17:00:00",
"2024-08-21T18:00:00",
"2024-08-21T19:00:00",
"2024-08-21T20:00:00",
"2024-08-21T21:00:00",
"2024-08-21T22:00:00",
"2024-08-21T23:00:00",
"2024-08-22T00:00:00",
"2024-08-22T01:00:00",
"2024-08-22T02:00:00",
"2024-08-22T03:00:00",
"2024-08-22T04:00:00",
"2024-08-22T05:00:00",
"2024-08-22T06:00:00",
"2024-08-22T07:00:00",
"2024-08-22T08:00:00",
"2024-08-22T09:00:00",
"2024-08-22T10:00:00",
"2024-08-22T11:00:00",
"2024-08-22T12:00:00",
"2024-08-22T13:00:00",
"2024-08-22T14:00:00",
"2024-08-22T15:00:00",
"2024-08-22T16:00:00",
"2024-08-22T17:00:00",
"2024-08-22T18:00:00",
"2024-08-22T19:00:00",
"2024-08-22T20:00:00",
"2024-08-22T21:00:00",
"2024-08-22T22:00:00",
"2024-08-22T23:00:00"
        ],
        "consumption-kwh": [
            1.343,
1.274,
1.126,
1.179,
1.468,
1.266,
1.111,
0.976,
1.198,
1.301,
1.417,
1.475,
2.434,
3.153,
3.795,
3.666,
4.048,
3.556,
4.358,
4.707,
4.402,
3.069,
2.632,
2.189,
1.824,
1.627,
1.544,
1.447,
1.42,
1.357,
1.553,
1.425,
1.987,
1.262,
1.524,
1.646,
2.983,
2.941,
2.924,
3.529,
4.003,
3.318,
3.762,
4.184,
3.597,
2.484,
2.135,
1.863,
1.428,
1.312,
1.223,
1.294,
1.352,
1.278,
1.173,
0.957,
1.312,
0.905,
1.73,
1.828,
1.812,
2.627,
2.827,
2.137,
2.279,
1.929,
2.939,
4.306,
3.312,
2.616,
1.946,
1.694,
1.451,
1.442,
1.232,
1.338,
1.234,
1.197,
1.31,
1.481,
1.225,
2.331,
2.546,
2.3,
2.316,
2.612,
2.543,
2.214,
2.322,
2.089,
2.484,
3.193,
2.66,
2.104,
1.816,
1.458,
1.194,
1.046,
1.087,
1.156,
1.104,
1.01,
1.162,
1.035,
1.287,
1.542,
1.545,
1.461,
1.63,
1.735,
2.3,
2.656,
2.527,
2.547,
3.684,
4.637,
4.577,
4.629,
3.561,
2.676,
2.065,
2.13,
1.392,
1.488,
1.259,
1.185,
1.132,
1.477,
1.183,
1.336,
1.591,
1.366,
1.548,
1.523,
1.641,
1.732,
1.516,
1.609,
2.319,
2.323,
2.068,
1.911,
1.687,
1.219,
1.047,
1.128,
1.008,
1.019,
1.214,
1.146,
1.122,
1.159,
1.064,
1.086,
1.152,
1.067,
1.334,
1.43,
1.321,
1.48,
4.176,
2.608,
3.142,
4.715,
4.763,
2.33,
2.091,
1.735,
1.22,
1.101,
1.09,
1.003,
1.047,
1.014,
1.211,
1.606,
1.407,
1.747,
1.298,
1.318,
1.313,
1.306,
1.409,
1.38,
1.82,
2.457,
1.779,
2.195,
2.112,
1.869,
1.374,
1.258,
1.318,
1.427,
1.168,
1.071,
1.118,
1.074,
1.29,
1.214,
1.127,
1.294,
1.051,
1.021,
1.026,
1.041,
1.308,
1.335,
1.509,
1.43,
2.143,
2.923,
2.361,
2.237,
1.676,
1.435,
1.412,
1.481,
1.326,
1.392,
1.176,
1.222,
1.35,
1.448,
1.619,
2.35,
2.064,
2.027,
2.384,
2.618,
3.828,
4.274,
3.791,
2.647,
2.744,
3.041,
2.39,
1.857,
1.547,
1.425,
1.353,
1.278,
1.136,
1.138,
1.117,
1.061,
1.042,
1.07,
1.061,
1.031,
1.36,
1.558,
1.593,
2.09,
1.534,
3.368,
3.656,
3.373,
2.872,
3.064,
2.325,
2.012,
1.678,
1.268,
1.418,
1.154,
0.998,
1.092,
0.998,
0.947,
1.211,
1.608,
1.41,
1.587,
1.328,
1.602,
1.43,
1.266,
1.516,
1.427,
1.692,
1.98,
2.181,
2.611,
2.443,
1.926,
1.356,
1.061,
1.176,
1.005,
1.079,
0.985,
1.006,
0.959,
1.144,
1.073,
1.155,
0.998,
0.947,
1.056,
1.21,
1.302,
1.541,
1.357,
1.516,
1.576,
2.14,
3.058,
2.854,
2.051,
1.374,
1.223,
1.193,
1.103,
1.039,
1.042,
1.132,
1.129,
1.081,
1.311,
1.796,
1.769,
1.927,
1.512,
1.461,
1.361,
1.579,
2.45,
2.135,
2.111,
2.492,
3.149,
2.088,
1.639,
0.967,
1.4,
1.077,
0.8,
0.77,
0.765,
0.617,
0.688,
0.798,
0.709,
0.74,
0.805,
0.765,
0.923,
1.064,
1.116,
1.109,
1.107,
1.443,
2.504,
2.619,
2.919,
2.163,
1.728,
1.227,
1.124,
1.012,
1.054,
0.835,
0.883,
0.833,
0.962,
1.383,
1.135,
1.335,
1.035,
0.815,
0.683,
0.722,
0.798,
0.687,
0.698,
0.652,
0.793,
1.337,
1.712,
1.523,
1.389,
1.297,
0.605,
0.546,
0.546,
0.532,
0.532,
0.392,
0.444,
0.458,
0.529,
0.515,
0.676,
0.598,
0.544,
0.659,
1.43,
1.071,
0.792,
1.089,
1.205,
1.466,
2.226,
1.914,
1.756,
1.417,
1.315,
0.912,
0.884,
0.896,
0.791,
0.832,
0.796,
0.639,
0.848,
0.768,
0.893,
1.34,
1.433,
1.113,
1.026,
0.975,
0.515,
0.611,
0.508,
0.477,
0.908,
0.951,
0.755,
0.742,
0.572,
0.607,
0.683,
0.538,
0.6,
0.519,
0.557,
0.505,
0.517,
0.498,
0.496,
0.623,
0.477,
0.5,
0.566,
0.503,
0.492,
0.547,
0.531,
0.538,
0.527,
0.501,
0.532,
0.524,
0.56,
0.569,
0.825,
0.628,
0.516,
0.55,
0.518,
0.514,
0.576,
0.518,
0.551,
0.513,
0.513,
0.553,
0.537,
0.494,
0.523,
0.544,
0.503,
0.526,
0.526,
0.566,
0.529,
0.501,
0.548,
0.521,
0.484,
0.541,
0.502,
0.518,
0.538,
0.536,
0.455,
0.478,
0.57,
0.482,
0.468,
0.578,
0.486,
0.462,
0.504,
0.529,
0.509,
0.531,
0.536,
0.547,
0.544,
0.504,
0.498,
0.552,
0.521,
0.5,
0.515,
0.511,
0.52,
0.531,
0.488,
0.52,
0.552,
0.467,
0.486,
0.526,
0.489,
0.506,
0.534,
0.504,
0.48,
0.532,
0.503,
0.529,
0.58,
0.538,
0.485
        ]
    }
}

Esempio di risposta REST

Per la richiesta di esempio viene restituita la seguente risposta.

{
  "model_id": "ibm/granite-ttm-512-96-r2",
  "created_at": "2024-11-18T15:27:09.312Z",
  "results": [
    {
      "consumption-kwh": [
        0.4290124177932739,
        0.4370437264442444,
        0.41718316078186035,
        0.38737165927886963,
        0.3955392837524414,
        0.3834899663925171,
        0.4044222831726074,
        0.42463231086730957,
        0.44074130058288574,
        0.467449426651001,
        0.47343355417251587,
        0.46828728914260864,
        0.5056831240653992,
        0.5539732575416565,
        0.5736022591590881,
        0.5886504054069519,
        0.6298335194587708,
        0.6267642974853516,
        0.7561632990837097,
        0.9204491376876831,
        0.8508875966072083,
        0.6771538257598877,
        0.5467019081115723,
        0.5154763460159302,
        0.4935459494590759,
        0.5060797333717346,
        0.4897392988204956,
        0.4714925289154053,
        0.48243576288223267,
        0.47915470600128174,
        0.5149773359298706,
        0.5487560629844666,
        0.579269528388977,
        0.614227831363678,
        0.6240154504776001,
        0.617196261882782,
        0.6460352540016174,
        0.6916171908378601,
        0.7112790942192078,
        0.7198072075843811,
        0.7407637238502502,
        0.7113152146339417,
        0.8186711668968201,
        0.9680589437484741,
        0.8914456963539124,
        0.7246706485748291,
        0.6004055738449097,
        0.5758842825889587,
        0.5714154243469238,
        0.5996972918510437,
        0.5872698426246643,
        0.5638936758041382,
        0.562979519367218,
        0.5384628176689148,
        0.5483810305595398,
        0.5603788495063782,
        0.5766607522964478,
        0.6183720231056213,
        0.6395336389541626,
        0.6367053985595703,
        0.6625468134880066,
        0.7083823084831238,
        0.7353737354278564,
        0.7451785206794739,
        0.7608164548873901,
        0.7376626133918762,
        0.8453754186630249,
        0.9998173117637634,
        0.9389885663986206,
        0.7839845418930054,
        0.6516207456588745,
        0.6147481799125671,
        0.5945475697517395,
        0.6096962690353394,
        0.5845052599906921,
        0.5627552270889282,
        0.5692415237426758,
        0.5633142590522766,
        0.5928394198417664,
        0.6212420463562012,
        0.6465822458267212,
        0.6908603310585022,
        0.7119491696357727,
        0.7027047276496887,
        0.7215489745140076,
        0.7617378234863281,
        0.7871705293655396,
        0.7981688976287842,
        0.8154757618904114,
        0.8133115768432617,
        0.9371682405471802,
        1.0931875705718994,
        1.0331265926361084,
        0.8758280873298645,
        0.7418281435966492,
        0.7037395238876343
      ],
      "date": [
        "2024-08-23T00:00:00",
        "2024-08-23T01:00:00",
        "2024-08-23T02:00:00",
        "2024-08-23T03:00:00",
        "2024-08-23T04:00:00",
        "2024-08-23T05:00:00",
        "2024-08-23T06:00:00",
        "2024-08-23T07:00:00",
        "2024-08-23T08:00:00",
        "2024-08-23T09:00:00",
        "2024-08-23T10:00:00",
        "2024-08-23T11:00:00",
        "2024-08-23T12:00:00",
        "2024-08-23T13:00:00",
        "2024-08-23T14:00:00",
        "2024-08-23T15:00:00",
        "2024-08-23T16:00:00",
        "2024-08-23T17:00:00",
        "2024-08-23T18:00:00",
        "2024-08-23T19:00:00",
        "2024-08-23T20:00:00",
        "2024-08-23T21:00:00",
        "2024-08-23T22:00:00",
        "2024-08-23T23:00:00",
        "2024-08-24T00:00:00",
        "2024-08-24T01:00:00",
        "2024-08-24T02:00:00",
        "2024-08-24T03:00:00",
        "2024-08-24T04:00:00",
        "2024-08-24T05:00:00",
        "2024-08-24T06:00:00",
        "2024-08-24T07:00:00",
        "2024-08-24T08:00:00",
        "2024-08-24T09:00:00",
        "2024-08-24T10:00:00",
        "2024-08-24T11:00:00",
        "2024-08-24T12:00:00",
        "2024-08-24T13:00:00",
        "2024-08-24T14:00:00",
        "2024-08-24T15:00:00",
        "2024-08-24T16:00:00",
        "2024-08-24T17:00:00",
        "2024-08-24T18:00:00",
        "2024-08-24T19:00:00",
        "2024-08-24T20:00:00",
        "2024-08-24T21:00:00",
        "2024-08-24T22:00:00",
        "2024-08-24T23:00:00",
        "2024-08-25T00:00:00",
        "2024-08-25T01:00:00",
        "2024-08-25T02:00:00",
        "2024-08-25T03:00:00",
        "2024-08-25T04:00:00",
        "2024-08-25T05:00:00",
        "2024-08-25T06:00:00",
        "2024-08-25T07:00:00",
        "2024-08-25T08:00:00",
        "2024-08-25T09:00:00",
        "2024-08-25T10:00:00",
        "2024-08-25T11:00:00",
        "2024-08-25T12:00:00",
        "2024-08-25T13:00:00",
        "2024-08-25T14:00:00",
        "2024-08-25T15:00:00",
        "2024-08-25T16:00:00",
        "2024-08-25T17:00:00",
        "2024-08-25T18:00:00",
        "2024-08-25T19:00:00",
        "2024-08-25T20:00:00",
        "2024-08-25T21:00:00",
        "2024-08-25T22:00:00",
        "2024-08-25T23:00:00",
        "2024-08-26T00:00:00",
        "2024-08-26T01:00:00",
        "2024-08-26T02:00:00",
        "2024-08-26T03:00:00",
        "2024-08-26T04:00:00",
        "2024-08-26T05:00:00",
        "2024-08-26T06:00:00",
        "2024-08-26T07:00:00",
        "2024-08-26T08:00:00",
        "2024-08-26T09:00:00",
        "2024-08-26T10:00:00",
        "2024-08-26T11:00:00",
        "2024-08-26T12:00:00",
        "2024-08-26T13:00:00",
        "2024-08-26T14:00:00",
        "2024-08-26T15:00:00",
        "2024-08-26T16:00:00",
        "2024-08-26T17:00:00",
        "2024-08-26T18:00:00",
        "2024-08-26T19:00:00",
        "2024-08-26T20:00:00",
        "2024-08-26T21:00:00",
        "2024-08-26T22:00:00",
        "2024-08-26T23:00:00"
      ],
      "ids": [
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A",
        "A"
      ]
    }
  ]
}

La tabella della schermata seguente mostra il consumo energetico previsto rispetto al consumo effettivo. La tabella indica che il modello è in grado di prevedere correttamente i picchi di consumo, anche se i numeri di consumo previsti dal modello sono più conservativi di quelli reali.

Mostra il consumo energetico orario effettivo rispetto a quello previsto.

Ulteriori informazioni

Argomento principale: Codificare soluzioni di IA generativa

Ricerca e risposta AI generativa
Queste risposte sono generate da un modello di lingua di grandi dimensioni in watsonx.ai basato sul contenuto della documentazione del prodotto. Ulteriori informazioni