I am developing an Android application that uses raw accelerometer data and I want to classify the data by using machine learning, i.e. Azure ML service. For example when device moves like a 1 in space, it should generate number 1 in text field specified in application. I decided to use Machine Learning to classify movements but I couldn't decide how to store data and send it to the machine learning service for training. For now, I am creating an SQLite table in application and add the X,Y,Z value of sensor each time sensor data gets changed. After that I am sending data to machine learning service but I have problem. The data only includes one movement for 1. How can I store multiple data for same movement and data for other movements -that will represent different numbers like 2, 3- and send them to the machine learning service?
@MuhammedKadirYücel, Based on my understanding, I think you want to send raw accelerometer data to Azure and store into some storage service for importing on Machine Learning service.
Per my experience, I think the best practice is that create a EventHub or IoTHub instance for receiving these raw accelerometer data.
Then create a Stream Analytics to transfer sensor data from EventHub or IoTHub to Azure Blob Storage.
Finally, you can import these data of the blob storage on Machine Learning service, please see https://azure.microsoft.com/en-us/documentation/articles/machine-learning-import-data-from-online-sources/.