I am a newbie in the field of AI/ML. I am trying to implement predictive analytics model on the data generated and collected every minute from a device with sensors.

I have two questions:

  1. What are various ML algorithms I can use to predict the number of people in a room given temperature, humidity, luminosity, and motion { 0 | 1 }?
  2. What other things can we predict using the above data from the sensor that is deployed in a closed room?


The device sends temperature, humidity, luminosity, and motion(yes/no) in real-time. I deployed this device in a closed room and started collecting data. Now I want to use this data to predict the number of people in the room using the data collected. I believe a multiple (linear/poly) regression model will help me in achieving this but, wanted to know if there are any other algorithms or any other use cases I can look into.

  • $\begingroup$ I expect there will be a lot of error margin (as I know people who can't sit still, and other who are extremely sedentary; some sweat a lot and other's not so much) but it's an extraordinarily interesting problem. Welcome to AI! $\endgroup$
    – DukeZhou
    Nov 21, 2017 at 22:30
  • 1
    $\begingroup$ What about sound? Humans make a lot of sound (breathing/heartbeat vibrations etc) depending on how good your sensors are and where they are located...? Just an idea... $\endgroup$
    – solarflare
    Nov 22, 2017 at 1:36
  • $\begingroup$ @DukeZhou one here, is trying to broaden the scope of the question. $\endgroup$
    – quintumnia
    Nov 23, 2017 at 9:45
  • $\begingroup$ @DenizC That's a great idea. Dogs will generally pant, or their collars give a little jingle. Wonder if the sensors can pick up things like heartbeat though... $\endgroup$
    – DukeZhou
    Nov 23, 2017 at 19:36
  • $\begingroup$ Google remote heartbeat sensors - military has been using them for years to detect heartbeats through walls. $\endgroup$
    – solarflare
    Nov 23, 2017 at 22:53


You must log in to answer this question.

Browse other questions tagged .