Hello guys, I am new here and I am currently in big dilemma. I have several questions that I would be very grateful to be answered.
our group’s building a smart home with Amazon echo, IoT and computer vision. I am the one who would be making the machine learning algorithm. However, we are in dilemma for our actual machine learning topic. Most of our ideas happened to be able to be solved by timer efficiently without any machine learning or highly effortless libraries. However, our current most promising idea is to use time series forecasting. We will use the prediction on what time a person leaves or enter their home to turn on or off certain machines.
ie. Based on a prediction, the air conditioner will turn on 5 minutes before the time the person predicted to go home. Based on a prediction, the man already left but he forgot to turn off the lights and AC. Then “after a week”, we will predict how much energy we saved compared to a timer based algorithm.
For our IoT, I am planning to use amazon. However, no lecturers are experienced in using AWS/Amazon Alexa. Developing an IoT + ML is very new to us so we do not know what to expect. Is there any limitation we should be aware under the context of my project?
Will this ML work and which algorithm should I use? If not, do you have any college-level deep learning ideas we can use?
Does amazon alexa/echo + IoT Tuya allow us easy integration of the ML frameworks on our IoT
if something is unclear, feel free to ask for any clarifications. Thanks, in advance.
Make sure you have enough data labeled for your application or if you can find it from similar projects thats even better but I doubt that.
I dont have much experience of ML and IOT and Alexa integration, it might be easy if AWS has frameworks ready for it. Otherwise in general its not an easy project.
a naive or linear prediction at least as a fair benchmark, as illustrated in this repo.
As a hint: besides the model resp. the (chained) algorithms, please consider that you will need many more components in your software system to realise an effective and satisfying operations such as pipeline monitoring, analytics, data cleaning, preprocessing, config, maintenance and other components to satisfy non-functional requirements as well as your specific MLOps concept etc., see also: Overview of ML Pipelines | Machine Learning | Google for Developers
Good thing is: with modern serverless infrastructure you can quickly build and validate your business hypothesis in a quite lean and scalable way.
I would be interested in your project. Feel free to keep us posted here if you like and please do not hesitate to ask for any advice or help.
hello, basically I am currently thinking of time series and linear regression on the project. While figuring out on the implementation, I decided to do something else. However, I reached a certain roadblock. As I promised, i made im giving an update on the development.
#prototype of value to time conversion
y1raw=
y1hour=
y1rawminute=
y1minute=
y1time=
for f1 in y1mean:
y1raw.append(f1/60)
y1raw = np.array(y1raw)
for f2 in y1raw:
y1hour.append(int(f2))
y1hour = np.array(y1hour)
y1raw = y1raw.reshape(1,7)
y1minute= np.round((y1raw - y1hour)*60)
print(y1raw)
print()
print(y1hour)
print()
print(y1minute)
for e1 in y1hour:
for e2 in y1minute:
y1time.append(e1,‘:’,e2)
I am trying to convert the value back from “number of minutes” that I used for calculating back to “time”. I was given a hint to check on “flat index to 2d index” but I dont understand np.unravel.
