Dear all

Greetings

Hope you all are safe and healthy. 

I would like to implement machine learning algorithms for my data which is about fatigue loading. Would you recommend which types of ML algorithms helps in this case the most? with obvious and complete predictive models. 

Thanks 
Cheers
Samira

------------------------------
Samira Gholizadeh
------------------------------

#GlobalAIandDataScience
#GlobalDataScience

Hello Samira! Hope you are ok as well.
Well, first of all I´ll assume that the data is idoneous and that it has been properly cleansed, because it really is more important than the model itself.
And in regards to the model, in order to select it you should consider which features you will use and which is your target. Could you give us with more details about the data?
For example, if it is about prediction then you should first see if you want to predict a numerical value (e.g: a fatigue accumulation metric) or if you want to predict the classification of an outcome (e.g: whether an athlete got injured or not).

------------------------------
[Marco] [Sánchez Sorondo]
[UBA]
------------------------------

Hello Marco

Thank you for your reply. 

The data has been collected in the lab and it is related to cyclic fatigue loading. I tried my best to cleanse and structure data. I am going to use predictive modeling to predicts some features on onset damages. I appreciate your advice Marco. Many thanks

Regards

------------------------------
Samira Gholizadeh
------------------------------

Original Message:
Sent: Mon August 10, 2020 03:38 PM
From: Marco Aurelio Sánchez Sorondo
Subject: Which types of ML works well in predictive modeling?

Hello Samira! Hope you are ok as well.
Well, first of all I´ll assume that the data is idoneous and that it has been properly cleansed, because it really is more important than the model itself.
And in regards to the model, in order to select it you should consider which features you will use and which is your target. Could you give us with more details about the data?
For example, if it is about prediction then you should first see if you want to predict a numerical value (e.g: a fatigue accumulation metric) or if you want to predict the classification of an outcome (e.g: whether an athlete got injured or not).

------------------------------
[Marco] [Sánchez Sorondo]
[UBA]
------------------------------

Okay, so... If the feature to predict is numerical then probably a random forest regressor or a gradient booster will do fine. If the feature to predict is categorical then a support vector machine would be useful, you could even get more from neural nets but I´m not the indicated to talk about them (still, haha). Try them and tell us how it goes!

------------------------------
[Marco] [Sánchez Sorondo]
[UBA]
------------------------------

Original Message:
Sent: Tue August 11, 2020 01:37 AM
From: Samira Gholizadeh
Subject: Which types of ML works well in predictive modeling?

Hello Marco

Thank you for your reply. 

The data has been collected in the lab and it is related to cyclic fatigue loading. I tried my best to cleanse and structure data. I am going to use predictive modeling to predicts some features on onset damages. I appreciate your advice Marco. Many thanks

Regards

------------------------------
Samira Gholizadeh
------------------------------


Original Message:
Sent: Mon August 10, 2020 03:38 PM
From: Marco Aurelio Sánchez Sorondo
Subject: Which types of ML works well in predictive modeling?

Hello Samira! Hope you are ok as well.
Well, first of all I´ll assume that the data is idoneous and that it has been properly cleansed, because it really is more important than the model itself.
And in regards to the model, in order to select it you should consider which features you will use and which is your target. Could you give us with more details about the data?
For example, if it is about prediction then you should first see if you want to predict a numerical value (e.g: a fatigue accumulation metric) or if you want to predict the classification of an outcome (e.g: whether an athlete got injured or not).

------------------------------
[Marco] [Sánchez Sorondo]
[UBA]

Thanks Marco. Let me try. I will keep you posted. Cheers

------------------------------
Samira Gholizadeh
------------------------------

Original Message:
Sent: Tue August 11, 2020 02:11 AM
From: Marco Aurelio Sánchez Sorondo
Subject: Which types of ML works well in predictive modeling?

Okay, so... If the feature to predict is numerical then probably a random forest regressor or a gradient booster will do fine. If the feature to predict is categorical then a support vector machine would be useful, you could even get more from neural nets but I´m not the indicated to talk about them (still, haha). Try them and tell us how it goes!

------------------------------
[Marco] [Sánchez Sorondo]
[UBA]
------------------------------


Original Message:
Sent: Tue August 11, 2020 01:37 AM
From: Samira Gholizadeh
Subject: Which types of ML works well in predictive modeling?

Hello Marco

Thank you for your reply. 

The data has been collected in the lab and it is related to cyclic fatigue loading. I tried my best to cleanse and structure data. I am going to use predictive modeling to predicts some features on onset damages. I appreciate your advice Marco. Many thanks

Regards

------------------------------
Samira Gholizadeh


Original Message:
Sent: Mon August 10, 2020 03:38 PM
From: Marco Aurelio Sánchez Sorondo
Subject: Which types of ML works well in predictive modeling?

Hello Samira! Hope you are ok as well.
Well, first of all I´ll assume that the data is idoneous and that it has been properly cleansed, because it really is more important than the model itself.
And in regards to the model, in order to select it you should consider which features you will use and which is your target. Could you give us with more details about the data?
For example, if it is about prediction then you should first see if you want to predict a numerical value (e.g: a fatigue accumulation metric) or if you want to predict the classification of an outcome (e.g: whether an athlete got injured or not).

------------------------------
[Marco] [Sánchez Sorondo]
[UBA]