Hfrxdjt

Air travel with refrigerated insulin

Why would five hundred and five be same as one?

Are Captain Marvel's powers affected by Thanos breaking the Tesseract and claiming the stone?

How much do grades matter for a future academia position?

Is it feasible to let a newcomer play the "Gandalf"-like figure I created for my campaign?

How to make a list of partial sums using forEach

How to leave product feedback on macOS?

Is there a RAID 0 Equivalent for RAM?

Grepping string, but include all non-blank lines following each grep match

What does "tick" mean in this sentence?

How to test the sharpness of a knife?

When is "ei" a diphthong?

How do you justify more code being written by following clean code practices?

How do I Interface a PS/2 Keyboard without Modern Techniques?

Why do Radio Buttons not fill the entire outer circle?

Is there anyway, I can have two passwords for my wi-fi

Giving feedback to someone without sounding prejudiced

Do I have to take mana from my deck or hand when tapping a dual land?

Check if object is null and return null

Review your own paper in Mathematics

Is there a reason to prefer HFS+ over APFS for disk images in High Sierra and/or Mojave?

Can I cause damage to electrical appliances by unplugging them when they are turned on?

"Oh no!" in Latin

Sound waves in different octaves

binary classification for time series data

How to merge monthly, daily and weekly data?Time series data: How I measure influence of new product sales on existing product sales (statistically)?What is a better approach for cross-validation with time-related predictorsHow do I find a repeating pattern of unknown length and start within a stringTime series finance -Correlation between a sector and MSCI ACWI returnsbinary classification for counting - estimating the error on counts due to error on prediction scoreDefining Input Shape for Time Series using LSTM in KerasDe-noising/removing measurement error from time series with very few observationsMultivariate Time Series Binary ClassificationBinary classification with time-series features

0

$begingroup$

I am a new to data science and I really appreciate any feedback on this problem.

I have a dataset with 450 subjects. A binary response (yes/no) is measured every week on each subject for a total of 104 weeks and the average proportion of events is 30%. Also, about 2000 features are measured each week from each subject to predict the weekly response. My questions are:

1- What is the best model to fit this data?
My concern about using machine learning models such as Random Forest or XGboost is that those models rely on the assumption that the analyzed data are independent and identically distributed (i.i.d.) which is not the case here because of the temporal correlation.

2- How do I split my data to train/test sets? For example, do I use only 350 subjects with 104w of data for training and 100 subjects with 104w for test set? Or use only 80 weeks of data for 450 subjects for training and 24w for 450 subjects as test set?

Thanks

classification time-series

share

asked 9 mins ago

AbuSalimAbuSalim

11

New contributor

AbuSalim is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.

$endgroup$

add a comment | 

0

$begingroup$

I am a new to data science and I really appreciate any feedback on this problem.

I have a dataset with 450 subjects. A binary response (yes/no) is measured every week on each subject for a total of 104 weeks and the average proportion of events is 30%. Also, about 2000 features are measured each week from each subject to predict the weekly response. My questions are:

1- What is the best model to fit this data?
My concern about using machine learning models such as Random Forest or XGboost is that those models rely on the assumption that the analyzed data are independent and identically distributed (i.i.d.) which is not the case here because of the temporal correlation.

2- How do I split my data to train/test sets? For example, do I use only 350 subjects with 104w of data for training and 100 subjects with 104w for test set? Or use only 80 weeks of data for 450 subjects for training and 24w for 450 subjects as test set?

Thanks

classification time-series

share

asked 9 mins ago

AbuSalimAbuSalim

11

New contributor

AbuSalim is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.

$endgroup$

add a comment | 

$begingroup$

I am a new to data science and I really appreciate any feedback on this problem.

I have a dataset with 450 subjects. A binary response (yes/no) is measured every week on each subject for a total of 104 weeks and the average proportion of events is 30%. Also, about 2000 features are measured each week from each subject to predict the weekly response. My questions are:

1- What is the best model to fit this data?
My concern about using machine learning models such as Random Forest or XGboost is that those models rely on the assumption that the analyzed data are independent and identically distributed (i.i.d.) which is not the case here because of the temporal correlation.

2- How do I split my data to train/test sets? For example, do I use only 350 subjects with 104w of data for training and 100 subjects with 104w for test set? Or use only 80 weeks of data for 450 subjects for training and 24w for 450 subjects as test set?

Thanks

classification time-series

share

asked 9 mins ago

AbuSalimAbuSalim

11

New contributor

AbuSalim is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.

$endgroup$

share

asked 9 mins ago

AbuSalimAbuSalim

11

New contributor

AbuSalim is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.

share

asked 9 mins ago

AbuSalimAbuSalim

11

New contributor

AbuSalim is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.

asked 9 mins ago

AbuSalimAbuSalim

11

New contributor

AbuSalim is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.

asked 9 mins ago

AbuSalimAbuSalim

11