Wild-Wav-Logo-White



Excel file upload

Upload
Excel File

Forecasting model



Forecasting model

Automatic Model Selection

Forecasting model



Forecasts

Generate
Forecasts



The web app has an interface with four steps:
  • Step 1: Choose an excel file with your data (format allowed: .xls or .xlsx). Preferably the file should have two columns: Date and Values (historical sales, visitors on a website, gold prices, stock prices, and many more). Download a sample file here.

  • Step 2: Choose the confidence interval that you like. The options provided are 99% CI, 95% CI, and 90% CI.

  • Step 3: Choose the forecast length of your wish. The options provided range from 1 month to 12 months (The forecasting model only provides monthly forecasting regardless of date format on the input file as of now).

  • Step 4: Click 'FORECAST'
Upon clicking the "forecast" button, the system shall retrieve the excel file and proceed to convert any dates that are not already in the YYYY-MM-DD format. In the event that the data is in a daily format, the system shall aggregate the values from the same month and compute their average. The resulting information shall then undergo a series of SARIMA models. The system shall select the model with the lowest Akaike Information Criterion (AIC) and apply said model to the data in order to generate forecasts along with their corresponding confidence intervals. The generated forecasts shall be presented in tabular format.
It is worth noting that the model yields more accurate results when working with either monthly or daily data that spans at least 4-6 years of historical data.

FAQs

Still have further queries? Some of the answers to these FAQs may help you clarify your doubts. If not, feel free to contact us via email or reach out to us using the feedback form.

  1. What is time series forecasting?

    2. Time series forecasting is a machine learning technique that uses information regarding historical values and associated patterns to predict future activity. Most often, this relates to trend analysis, cyclical fluctuation analysis, and issues of seasonality.

  2. How much does it cost to Forecast?

    3. Generating forecasts on the Auto-Forecasting web application is free of cost. Should you wish to save the generated data, you may download or print the forecasts.

  3. Does my data get stored within the application?

    4. Auto-Forecasting is designed such that the user's data does not get stored on the server. After the forecasting process is complete, the uploaded data gets automatically deleted.

  4. The application is taking time to forecast. Why?

    5. The Auto-Forecasting application uses machine-learning to recognize and opt for the best model based on the data that is provided. This process takes time due to the high computational complexity. We are working on making the system more faster and effecient.