₦airaland Forum

I recently came across machine learning and i have some questions that has been bothering me. I'd be glad if you kindly help me out.


How do Machine learners use their alogrithms/model? E.G... A machine learner creates a model that can predicts house prices in Ojuelegba; will he create a software/app to implement the code or he just leaves it raw?. I'm new to all these, so forgive my misuse of terminologies.


Please help me out to explain this field because i want to major in it. The explanations online ain't helpful. Help Me Out Guys.

Hello.
In data science there’s something called “train test split”, where you split your sample data into a given ration(maybe 70%). And train your model again the training data.
Once you have trained your data and tested is accuracy with the test data. You can serialize and save the model and use subsequently for new predications.
As long as you are using the model for similar predictions, for instance, house prices like you said, you can always reuse the model. Write the model meta data into a file, or database.

Realoaac:
Hello.
In data science there’s something called “train test split”, where you split your sample data into a given ration(maybe 70%). And train your model again the training data.
Once you have trained your data and tested is accuracy with the test data. You can serialize and save the model and use subsequently for new predications.
As long as you are using the model for similar predictions, for instance, house prices like you said, you can always reuse the model. Write the model meta data into a file, or database.

Thanks for answering him bro.

You may not have it as a software and if you have it as a software then it will be dedicated for that problem only. So you can always save it as a file or database as he has written.

Other experts will come in to give more insight.

Guy if you dey seek help on machine learning from nairaland, you're on a loooooooooong thing

Realoaac:
Hello.
In data science there’s something called “train test split”, where you split your sample data into a given ration(maybe 70%). And train your model again the training data.
Once you have trained your data and tested is accuracy with the test data. You can serialize and save the model and use subsequently for new predications.
As long as you are using the model for similar predictions, for instance, house prices like you said, you can always reuse the model. Write the model meta data into a file, or database.

Thanks man, i really appreciate this.

shams040:
Thanks for answering him bro.

You may not have it as a software and if you have it as a software then it will be dedicated for that problem only. So you can always save it as a file or database as he has written.

Other experts will come in to give more insight.

Thanks man, that was what i wanted to hear. I see you are into data analysis, what Programming Lang can you recommend and what mathematical topic should i master?

jerflakes:
Guy if you dey seek help on machine learning from nairaland, you're on a loooooooooong thing

why na? Do we only have wordpress developers on nairaland?

Aquactic:
Thanks man, that was what i wanted to hear. I see you are into data analysis, what Programming Lang can you recommend and what mathematical topic should i master?

Use R for now but prior to this time I use Spss. I also use GIS for spatial mapping and Analysis

Aquactic:
I recently came across machine learning and i have some questions that has been bothering me. I'd be glad if you kindly help me out.


How do Machine learners use their alogrithms/model? E.G... A machine learner creates a model that can predicts house prices in Ojuelegba; will he create a software/app to implement the code or he just leaves it raw?. I'm new to all these, so forgive my misuse of terminologies.


Please help me out to explain this field because i want to major in it. The explanations online ain't helpful. Help Me Out Guys.

Very good question.
So you have trained data and tested model for prediction. deploying your model into production may vary across use cases. Your model can be deployed anywhere after you have saved it and sure about what your model is intended to do. Since you are new to ML/AI, there is still quite a lot of learning curve.
take the coursera courses on Machine learning and deep learning, practice, practice, practice.

email me to discuss further and see if there is a way I can help: dunjoye4real@gmail.com, www.linkedin.com/in/olarrysh