Feature engineering suggestion required The 2019 Stack Overflow Developer Survey Results Are InFeature Extraction Technique - Summarizing a Sequence of DataPrepping Data For Usage ClusteringGround-truth and feature extraction for predictive modellingHow to use neural network's hidden layer output for feature engineering?Fix missing data by adding another feature instead of using the mean?What are best practices for collaborative feature engineering?How would knowing spammers email address improve spam detection algorithms?Is this a good practice of feature engineering?How do I develop a system to Recommend a marketing channel using data science?Cant get LSTM model to give required predictions

What could be the right powersource for 15 seconds lifespan disposable giant chainsaw?

Are there incongruent pythagorean triangles with the same perimeter and same area?

The difference between dialogue marks

Why was M87 targetted for the Event Horizon Telescope instead of Sagittarius A*?

Are there any other methods to apply to solving simultaneous equations?

Button changing it's text & action. Good or terrible?

Is a "Democratic" Oligarchy-Style System Possible?

Does a dangling wire really electrocute me if I'm standing in water?

Why is the Constellation's nose gear so long?

For what reasons would an animal species NOT cross a *horizontal* land bridge?

slides for 30min~1hr skype tenure track application interview

Why do UK politicians seemingly ignore opinion polls on Brexit?

Right tool to dig six foot holes?

Are spiders unable to hurt humans, especially very small spiders?

Which Sci-Fi work first showed weapon of galactic-scale mass destruction?

Why isn't airport relocation done gradually?

Why not take a picture of a closer black hole?

Why do we hear so much about the Trump administration deciding to impose and then remove tariffs?

One word riddle: Vowel in the middle

What do hard-Brexiteers want with respect to the Irish border?

What is the meaning of the verb "bear" in this context?

Did 3000BC Egyptians use meteoric iron weapons?

Do these rules for Critical Successes and Critical Failures seem fair?

Pokemon Turn Based battle (Python)



Feature engineering suggestion required



The 2019 Stack Overflow Developer Survey Results Are InFeature Extraction Technique - Summarizing a Sequence of DataPrepping Data For Usage ClusteringGround-truth and feature extraction for predictive modellingHow to use neural network's hidden layer output for feature engineering?Fix missing data by adding another feature instead of using the mean?What are best practices for collaborative feature engineering?How would knowing spammers email address improve spam detection algorithms?Is this a good practice of feature engineering?How do I develop a system to Recommend a marketing channel using data science?Cant get LSTM model to give required predictions










4












$begingroup$


I am having a problem during feature engineering. Looking for some suggestions. Problem statement: I have usage data of multiple customers for 3 days. Some have just 1 day usage some 2 and some 3. Data is related to number of emails sent / contacts added on each day etc.



I am converting this time series data to column-wise ie., number of emails sent by a customer on day1 as one feature, number of emails sent by a customer on day2 as one feature and so on. But problem is that, the usage can be of either increasing order or decreasing order for different customers.



ie., example 1: customer 'A' --> 'number of emails sent on 1st . day' = 100 . ' number of emails sent on 2nd day'=0



example 2: customer 'B' --> 'number of emails sent on 1st . day' = 0 . ' number of emails sent on 2nd day'=100



example 3: customer 'C' --> 'number of emails sent on 1st . day' = 0 . ' number of emails sent on 2nd day'=0



example 4: customer 'D' --> 'number of emails sent on 1st . day' = 100 . ' number of emails sent on 2nd day'=100



In the first two cases => My new feature will have "-100" and "100" as values. Which I guess is good for differentiating. But the problem arises for 3rd and 4th columns when the new feature value will be "0" in both scenarios Can anyone suggest a way to handle this.



One way to handle this:



I can add "No change" in those scenarios, but I am confused about one thing. If I do that, I will have to make the new feature as categorical, which is not ideal as the other values will be continuous.



Instead, I can have absolute values in the new feature and indicate the trend as "+1" or increasing "-1" for decreasing "no change" for no change and "0" if both the values have been "0". Would that be a good approach though?



The end goal is to predict if a user would continue using the application or not. So it basically would be a two-class model. And I would want to capture even the scale of usage i.e., "A user sending 100 emails every day" should be different from "B user sending 10000 emails every day"










share|improve this question









New contributor




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







$endgroup$







  • 1




    $begingroup$
    could you explain a bit better what are you trying to predict? Your question is pretty well explained but the kind of model you plan do train might give some of us better ideas.
    $endgroup$
    – Pedro Henrique Monforte
    2 hours ago










  • $begingroup$
    I would want to predict if a user would continue using the application or not. So it basically would be a two-class model. Does that answer?
    $endgroup$
    – SSuram
    1 hour ago










  • $begingroup$
    Yes, just add it to your question and it will be perfect
    $endgroup$
    – Pedro Henrique Monforte
    1 hour ago















4












$begingroup$


I am having a problem during feature engineering. Looking for some suggestions. Problem statement: I have usage data of multiple customers for 3 days. Some have just 1 day usage some 2 and some 3. Data is related to number of emails sent / contacts added on each day etc.



I am converting this time series data to column-wise ie., number of emails sent by a customer on day1 as one feature, number of emails sent by a customer on day2 as one feature and so on. But problem is that, the usage can be of either increasing order or decreasing order for different customers.



ie., example 1: customer 'A' --> 'number of emails sent on 1st . day' = 100 . ' number of emails sent on 2nd day'=0



example 2: customer 'B' --> 'number of emails sent on 1st . day' = 0 . ' number of emails sent on 2nd day'=100



example 3: customer 'C' --> 'number of emails sent on 1st . day' = 0 . ' number of emails sent on 2nd day'=0



example 4: customer 'D' --> 'number of emails sent on 1st . day' = 100 . ' number of emails sent on 2nd day'=100



In the first two cases => My new feature will have "-100" and "100" as values. Which I guess is good for differentiating. But the problem arises for 3rd and 4th columns when the new feature value will be "0" in both scenarios Can anyone suggest a way to handle this.



One way to handle this:



I can add "No change" in those scenarios, but I am confused about one thing. If I do that, I will have to make the new feature as categorical, which is not ideal as the other values will be continuous.



Instead, I can have absolute values in the new feature and indicate the trend as "+1" or increasing "-1" for decreasing "no change" for no change and "0" if both the values have been "0". Would that be a good approach though?



The end goal is to predict if a user would continue using the application or not. So it basically would be a two-class model. And I would want to capture even the scale of usage i.e., "A user sending 100 emails every day" should be different from "B user sending 10000 emails every day"










share|improve this question









New contributor




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







$endgroup$







  • 1




    $begingroup$
    could you explain a bit better what are you trying to predict? Your question is pretty well explained but the kind of model you plan do train might give some of us better ideas.
    $endgroup$
    – Pedro Henrique Monforte
    2 hours ago










  • $begingroup$
    I would want to predict if a user would continue using the application or not. So it basically would be a two-class model. Does that answer?
    $endgroup$
    – SSuram
    1 hour ago










  • $begingroup$
    Yes, just add it to your question and it will be perfect
    $endgroup$
    – Pedro Henrique Monforte
    1 hour ago













4












4








4





$begingroup$


I am having a problem during feature engineering. Looking for some suggestions. Problem statement: I have usage data of multiple customers for 3 days. Some have just 1 day usage some 2 and some 3. Data is related to number of emails sent / contacts added on each day etc.



I am converting this time series data to column-wise ie., number of emails sent by a customer on day1 as one feature, number of emails sent by a customer on day2 as one feature and so on. But problem is that, the usage can be of either increasing order or decreasing order for different customers.



ie., example 1: customer 'A' --> 'number of emails sent on 1st . day' = 100 . ' number of emails sent on 2nd day'=0



example 2: customer 'B' --> 'number of emails sent on 1st . day' = 0 . ' number of emails sent on 2nd day'=100



example 3: customer 'C' --> 'number of emails sent on 1st . day' = 0 . ' number of emails sent on 2nd day'=0



example 4: customer 'D' --> 'number of emails sent on 1st . day' = 100 . ' number of emails sent on 2nd day'=100



In the first two cases => My new feature will have "-100" and "100" as values. Which I guess is good for differentiating. But the problem arises for 3rd and 4th columns when the new feature value will be "0" in both scenarios Can anyone suggest a way to handle this.



One way to handle this:



I can add "No change" in those scenarios, but I am confused about one thing. If I do that, I will have to make the new feature as categorical, which is not ideal as the other values will be continuous.



Instead, I can have absolute values in the new feature and indicate the trend as "+1" or increasing "-1" for decreasing "no change" for no change and "0" if both the values have been "0". Would that be a good approach though?



The end goal is to predict if a user would continue using the application or not. So it basically would be a two-class model. And I would want to capture even the scale of usage i.e., "A user sending 100 emails every day" should be different from "B user sending 10000 emails every day"










share|improve this question









New contributor




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







$endgroup$




I am having a problem during feature engineering. Looking for some suggestions. Problem statement: I have usage data of multiple customers for 3 days. Some have just 1 day usage some 2 and some 3. Data is related to number of emails sent / contacts added on each day etc.



I am converting this time series data to column-wise ie., number of emails sent by a customer on day1 as one feature, number of emails sent by a customer on day2 as one feature and so on. But problem is that, the usage can be of either increasing order or decreasing order for different customers.



ie., example 1: customer 'A' --> 'number of emails sent on 1st . day' = 100 . ' number of emails sent on 2nd day'=0



example 2: customer 'B' --> 'number of emails sent on 1st . day' = 0 . ' number of emails sent on 2nd day'=100



example 3: customer 'C' --> 'number of emails sent on 1st . day' = 0 . ' number of emails sent on 2nd day'=0



example 4: customer 'D' --> 'number of emails sent on 1st . day' = 100 . ' number of emails sent on 2nd day'=100



In the first two cases => My new feature will have "-100" and "100" as values. Which I guess is good for differentiating. But the problem arises for 3rd and 4th columns when the new feature value will be "0" in both scenarios Can anyone suggest a way to handle this.



One way to handle this:



I can add "No change" in those scenarios, but I am confused about one thing. If I do that, I will have to make the new feature as categorical, which is not ideal as the other values will be continuous.



Instead, I can have absolute values in the new feature and indicate the trend as "+1" or increasing "-1" for decreasing "no change" for no change and "0" if both the values have been "0". Would that be a good approach though?



The end goal is to predict if a user would continue using the application or not. So it basically would be a two-class model. And I would want to capture even the scale of usage i.e., "A user sending 100 emails every day" should be different from "B user sending 10000 emails every day"







machine-learning feature-engineering data-science-model






share|improve this question









New contributor




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











share|improve this question









New contributor




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









share|improve this question




share|improve this question








edited 1 hour ago







SSuram













New contributor




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









asked 2 hours ago









SSuramSSuram

214




214




New contributor




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





New contributor





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






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







  • 1




    $begingroup$
    could you explain a bit better what are you trying to predict? Your question is pretty well explained but the kind of model you plan do train might give some of us better ideas.
    $endgroup$
    – Pedro Henrique Monforte
    2 hours ago










  • $begingroup$
    I would want to predict if a user would continue using the application or not. So it basically would be a two-class model. Does that answer?
    $endgroup$
    – SSuram
    1 hour ago










  • $begingroup$
    Yes, just add it to your question and it will be perfect
    $endgroup$
    – Pedro Henrique Monforte
    1 hour ago












  • 1




    $begingroup$
    could you explain a bit better what are you trying to predict? Your question is pretty well explained but the kind of model you plan do train might give some of us better ideas.
    $endgroup$
    – Pedro Henrique Monforte
    2 hours ago










  • $begingroup$
    I would want to predict if a user would continue using the application or not. So it basically would be a two-class model. Does that answer?
    $endgroup$
    – SSuram
    1 hour ago










  • $begingroup$
    Yes, just add it to your question and it will be perfect
    $endgroup$
    – Pedro Henrique Monforte
    1 hour ago







1




1




$begingroup$
could you explain a bit better what are you trying to predict? Your question is pretty well explained but the kind of model you plan do train might give some of us better ideas.
$endgroup$
– Pedro Henrique Monforte
2 hours ago




$begingroup$
could you explain a bit better what are you trying to predict? Your question is pretty well explained but the kind of model you plan do train might give some of us better ideas.
$endgroup$
– Pedro Henrique Monforte
2 hours ago












$begingroup$
I would want to predict if a user would continue using the application or not. So it basically would be a two-class model. Does that answer?
$endgroup$
– SSuram
1 hour ago




$begingroup$
I would want to predict if a user would continue using the application or not. So it basically would be a two-class model. Does that answer?
$endgroup$
– SSuram
1 hour ago












$begingroup$
Yes, just add it to your question and it will be perfect
$endgroup$
– Pedro Henrique Monforte
1 hour ago




$begingroup$
Yes, just add it to your question and it will be perfect
$endgroup$
– Pedro Henrique Monforte
1 hour ago










1 Answer
1






active

oldest

votes


















2












$begingroup$

Well, you want to identify change in usage you could try something like:



$$ f(day_1,day2) = fracday_2-day_1 + delta times Biggr|Biggr|fracday_2+day_1(day_2+day_1+1)(day_2-day_1+1)Biggl|Biggl| $$



where $delta$ is the eps of your machine (minimum value needed to be summed to differ it from other floats)



that will give you
$$f(100,0) approx -98.02$$
$$f(0,100) = 100$$
$$f(100,100) approx 0.995$$
$$f(0,0) = 0$$



You can look at my experiment here



This will map all non-changes from $[0,1]$ where $f(0,0)$ maps to $0$ and $f(infty,infty)$ maps to $1$



Where is it from? Just tuned the function manually. But I think this might suffice for your application



Explaining the ideia



You want to have a feature that packs a lot of information:
- Is the usage greater than zero?
- Is it increasing or decreasing?
- If it is stalled, how much is the usage?



Well, your usage vary in integer values so you can map the entire non-changing but above 0 case to a previously non-used interval.



The function above will map in $[0,1]$ all non-changing possibilities, in a exponential kind of way ($a^(-frac1usage)$) also you can extract the actual value from positive changes and the approximate value for negative change (been a better approximation when the drop is high)



This is not the perfect scenario but it is the maximum information I could compress into 1 variable with little loss.






share|improve this answer











$endgroup$












  • $begingroup$
    I am not sure if it would answer --- "'''And I would want to capture even the scale of usage i.e., "A user sending 100 emails every day" should be different from "B user sending 10000 emails every day ""''---- part of the question. Could you please explain?
    $endgroup$
    – SSuram
    1 hour ago










  • $begingroup$
    What would you say about adding the below info to it f = (((d2-d1+eps)/abs(d2-d1+eps))*abs((d2+d1)/(d1+d2+1)*(d2-d1+1)))*(d2/1000)*(d1/1000) where "1000"-- would be max(usage).
    $endgroup$
    – SSuram
    54 mins ago











  • $begingroup$
    that will actually return zero for near every case
    $endgroup$
    – Pedro Henrique Monforte
    43 mins ago











Your Answer





StackExchange.ifUsing("editor", function ()
return StackExchange.using("mathjaxEditing", function ()
StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix)
StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
);
);
, "mathjax-editing");

StackExchange.ready(function()
var channelOptions =
tags: "".split(" "),
id: "557"
;
initTagRenderer("".split(" "), "".split(" "), channelOptions);

StackExchange.using("externalEditor", function()
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled)
StackExchange.using("snippets", function()
createEditor();
);

else
createEditor();

);

function createEditor()
StackExchange.prepareEditor(
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: false,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: null,
bindNavPrevention: true,
postfix: "",
imageUploader:
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
,
onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
);



);






SSuram is a new contributor. Be nice, and check out our Code of Conduct.









draft saved

draft discarded


















StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f49088%2ffeature-engineering-suggestion-required%23new-answer', 'question_page');

);

Post as a guest















Required, but never shown

























1 Answer
1






active

oldest

votes








1 Answer
1






active

oldest

votes









active

oldest

votes






active

oldest

votes









2












$begingroup$

Well, you want to identify change in usage you could try something like:



$$ f(day_1,day2) = fracday_2-day_1 + delta times Biggr|Biggr|fracday_2+day_1(day_2+day_1+1)(day_2-day_1+1)Biggl|Biggl| $$



where $delta$ is the eps of your machine (minimum value needed to be summed to differ it from other floats)



that will give you
$$f(100,0) approx -98.02$$
$$f(0,100) = 100$$
$$f(100,100) approx 0.995$$
$$f(0,0) = 0$$



You can look at my experiment here



This will map all non-changes from $[0,1]$ where $f(0,0)$ maps to $0$ and $f(infty,infty)$ maps to $1$



Where is it from? Just tuned the function manually. But I think this might suffice for your application



Explaining the ideia



You want to have a feature that packs a lot of information:
- Is the usage greater than zero?
- Is it increasing or decreasing?
- If it is stalled, how much is the usage?



Well, your usage vary in integer values so you can map the entire non-changing but above 0 case to a previously non-used interval.



The function above will map in $[0,1]$ all non-changing possibilities, in a exponential kind of way ($a^(-frac1usage)$) also you can extract the actual value from positive changes and the approximate value for negative change (been a better approximation when the drop is high)



This is not the perfect scenario but it is the maximum information I could compress into 1 variable with little loss.






share|improve this answer











$endgroup$












  • $begingroup$
    I am not sure if it would answer --- "'''And I would want to capture even the scale of usage i.e., "A user sending 100 emails every day" should be different from "B user sending 10000 emails every day ""''---- part of the question. Could you please explain?
    $endgroup$
    – SSuram
    1 hour ago










  • $begingroup$
    What would you say about adding the below info to it f = (((d2-d1+eps)/abs(d2-d1+eps))*abs((d2+d1)/(d1+d2+1)*(d2-d1+1)))*(d2/1000)*(d1/1000) where "1000"-- would be max(usage).
    $endgroup$
    – SSuram
    54 mins ago











  • $begingroup$
    that will actually return zero for near every case
    $endgroup$
    – Pedro Henrique Monforte
    43 mins ago















2












$begingroup$

Well, you want to identify change in usage you could try something like:



$$ f(day_1,day2) = fracday_2-day_1 + delta times Biggr|Biggr|fracday_2+day_1(day_2+day_1+1)(day_2-day_1+1)Biggl|Biggl| $$



where $delta$ is the eps of your machine (minimum value needed to be summed to differ it from other floats)



that will give you
$$f(100,0) approx -98.02$$
$$f(0,100) = 100$$
$$f(100,100) approx 0.995$$
$$f(0,0) = 0$$



You can look at my experiment here



This will map all non-changes from $[0,1]$ where $f(0,0)$ maps to $0$ and $f(infty,infty)$ maps to $1$



Where is it from? Just tuned the function manually. But I think this might suffice for your application



Explaining the ideia



You want to have a feature that packs a lot of information:
- Is the usage greater than zero?
- Is it increasing or decreasing?
- If it is stalled, how much is the usage?



Well, your usage vary in integer values so you can map the entire non-changing but above 0 case to a previously non-used interval.



The function above will map in $[0,1]$ all non-changing possibilities, in a exponential kind of way ($a^(-frac1usage)$) also you can extract the actual value from positive changes and the approximate value for negative change (been a better approximation when the drop is high)



This is not the perfect scenario but it is the maximum information I could compress into 1 variable with little loss.






share|improve this answer











$endgroup$












  • $begingroup$
    I am not sure if it would answer --- "'''And I would want to capture even the scale of usage i.e., "A user sending 100 emails every day" should be different from "B user sending 10000 emails every day ""''---- part of the question. Could you please explain?
    $endgroup$
    – SSuram
    1 hour ago










  • $begingroup$
    What would you say about adding the below info to it f = (((d2-d1+eps)/abs(d2-d1+eps))*abs((d2+d1)/(d1+d2+1)*(d2-d1+1)))*(d2/1000)*(d1/1000) where "1000"-- would be max(usage).
    $endgroup$
    – SSuram
    54 mins ago











  • $begingroup$
    that will actually return zero for near every case
    $endgroup$
    – Pedro Henrique Monforte
    43 mins ago













2












2








2





$begingroup$

Well, you want to identify change in usage you could try something like:



$$ f(day_1,day2) = fracday_2-day_1 + delta times Biggr|Biggr|fracday_2+day_1(day_2+day_1+1)(day_2-day_1+1)Biggl|Biggl| $$



where $delta$ is the eps of your machine (minimum value needed to be summed to differ it from other floats)



that will give you
$$f(100,0) approx -98.02$$
$$f(0,100) = 100$$
$$f(100,100) approx 0.995$$
$$f(0,0) = 0$$



You can look at my experiment here



This will map all non-changes from $[0,1]$ where $f(0,0)$ maps to $0$ and $f(infty,infty)$ maps to $1$



Where is it from? Just tuned the function manually. But I think this might suffice for your application



Explaining the ideia



You want to have a feature that packs a lot of information:
- Is the usage greater than zero?
- Is it increasing or decreasing?
- If it is stalled, how much is the usage?



Well, your usage vary in integer values so you can map the entire non-changing but above 0 case to a previously non-used interval.



The function above will map in $[0,1]$ all non-changing possibilities, in a exponential kind of way ($a^(-frac1usage)$) also you can extract the actual value from positive changes and the approximate value for negative change (been a better approximation when the drop is high)



This is not the perfect scenario but it is the maximum information I could compress into 1 variable with little loss.






share|improve this answer











$endgroup$



Well, you want to identify change in usage you could try something like:



$$ f(day_1,day2) = fracday_2-day_1 + delta times Biggr|Biggr|fracday_2+day_1(day_2+day_1+1)(day_2-day_1+1)Biggl|Biggl| $$



where $delta$ is the eps of your machine (minimum value needed to be summed to differ it from other floats)



that will give you
$$f(100,0) approx -98.02$$
$$f(0,100) = 100$$
$$f(100,100) approx 0.995$$
$$f(0,0) = 0$$



You can look at my experiment here



This will map all non-changes from $[0,1]$ where $f(0,0)$ maps to $0$ and $f(infty,infty)$ maps to $1$



Where is it from? Just tuned the function manually. But I think this might suffice for your application



Explaining the ideia



You want to have a feature that packs a lot of information:
- Is the usage greater than zero?
- Is it increasing or decreasing?
- If it is stalled, how much is the usage?



Well, your usage vary in integer values so you can map the entire non-changing but above 0 case to a previously non-used interval.



The function above will map in $[0,1]$ all non-changing possibilities, in a exponential kind of way ($a^(-frac1usage)$) also you can extract the actual value from positive changes and the approximate value for negative change (been a better approximation when the drop is high)



This is not the perfect scenario but it is the maximum information I could compress into 1 variable with little loss.







share|improve this answer














share|improve this answer



share|improve this answer








edited 46 mins ago


























community wiki





2 revs
Pedro Henrique Monforte












  • $begingroup$
    I am not sure if it would answer --- "'''And I would want to capture even the scale of usage i.e., "A user sending 100 emails every day" should be different from "B user sending 10000 emails every day ""''---- part of the question. Could you please explain?
    $endgroup$
    – SSuram
    1 hour ago










  • $begingroup$
    What would you say about adding the below info to it f = (((d2-d1+eps)/abs(d2-d1+eps))*abs((d2+d1)/(d1+d2+1)*(d2-d1+1)))*(d2/1000)*(d1/1000) where "1000"-- would be max(usage).
    $endgroup$
    – SSuram
    54 mins ago











  • $begingroup$
    that will actually return zero for near every case
    $endgroup$
    – Pedro Henrique Monforte
    43 mins ago
















  • $begingroup$
    I am not sure if it would answer --- "'''And I would want to capture even the scale of usage i.e., "A user sending 100 emails every day" should be different from "B user sending 10000 emails every day ""''---- part of the question. Could you please explain?
    $endgroup$
    – SSuram
    1 hour ago










  • $begingroup$
    What would you say about adding the below info to it f = (((d2-d1+eps)/abs(d2-d1+eps))*abs((d2+d1)/(d1+d2+1)*(d2-d1+1)))*(d2/1000)*(d1/1000) where "1000"-- would be max(usage).
    $endgroup$
    – SSuram
    54 mins ago











  • $begingroup$
    that will actually return zero for near every case
    $endgroup$
    – Pedro Henrique Monforte
    43 mins ago















$begingroup$
I am not sure if it would answer --- "'''And I would want to capture even the scale of usage i.e., "A user sending 100 emails every day" should be different from "B user sending 10000 emails every day ""''---- part of the question. Could you please explain?
$endgroup$
– SSuram
1 hour ago




$begingroup$
I am not sure if it would answer --- "'''And I would want to capture even the scale of usage i.e., "A user sending 100 emails every day" should be different from "B user sending 10000 emails every day ""''---- part of the question. Could you please explain?
$endgroup$
– SSuram
1 hour ago












$begingroup$
What would you say about adding the below info to it f = (((d2-d1+eps)/abs(d2-d1+eps))*abs((d2+d1)/(d1+d2+1)*(d2-d1+1)))*(d2/1000)*(d1/1000) where "1000"-- would be max(usage).
$endgroup$
– SSuram
54 mins ago





$begingroup$
What would you say about adding the below info to it f = (((d2-d1+eps)/abs(d2-d1+eps))*abs((d2+d1)/(d1+d2+1)*(d2-d1+1)))*(d2/1000)*(d1/1000) where "1000"-- would be max(usage).
$endgroup$
– SSuram
54 mins ago













$begingroup$
that will actually return zero for near every case
$endgroup$
– Pedro Henrique Monforte
43 mins ago




$begingroup$
that will actually return zero for near every case
$endgroup$
– Pedro Henrique Monforte
43 mins ago










SSuram is a new contributor. Be nice, and check out our Code of Conduct.









draft saved

draft discarded


















SSuram is a new contributor. Be nice, and check out our Code of Conduct.












SSuram is a new contributor. Be nice, and check out our Code of Conduct.











SSuram is a new contributor. Be nice, and check out our Code of Conduct.














Thanks for contributing an answer to Data Science Stack Exchange!


  • Please be sure to answer the question. Provide details and share your research!

But avoid …


  • Asking for help, clarification, or responding to other answers.

  • Making statements based on opinion; back them up with references or personal experience.

Use MathJax to format equations. MathJax reference.


To learn more, see our tips on writing great answers.




draft saved


draft discarded














StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f49088%2ffeature-engineering-suggestion-required%23new-answer', 'question_page');

);

Post as a guest















Required, but never shown





















































Required, but never shown














Required, but never shown












Required, but never shown







Required, but never shown

































Required, but never shown














Required, but never shown












Required, but never shown







Required, but never shown







Popular posts from this blog

Is it wise to focus on putting odd beats on left when playing double bass drums?What does the tie across other different notes mean?How to improve right hand picking/strumming technique for electric and acoustic guitar?How fast is the BPM should you start to use double bass pedal?Alternative or more advanced methods for counting rhythmsLost direction in practicing drumsIs there any way at all to tell if these measures are 6/4 vs. 3/2?Why is alberti bass so tiring?How does 'meter' differ from 'rhythm' in music?Rhythm question about piano sheet notationWhy is Bach's Cello Suite 1 written in 16th notes

map list to bin numbersquantilization (if that is a word)Map efficiently over duplicates in listGenerating a list of cubefree numbersIs there an equivalent to MATLAB's linspace?Convert a list of hexadecimal numbers to decimalTaking one list Mod a second listlist of items and group of alternative itemsHow find numbers in this list of inequalities?Selecting list entries with a True False index list of similar lengthReplace element in array by checking condition in another listAttempting to fill a table with the number of elements in each bin and make a table with the elements in the bins?

History of India Contents Prehistoric era (until c. 3300 BCE) Bronze Age – "First urbanisation" (c. 3300 – c. 1800 BCE) Climate change, de-urbanisation, and Indo-Aryan migrations (c.1800 – 1500 BCE) Iron Age - Vedic period (c. 1500 – c. 600 BCE) "Second urbanisation" (c. 600 – c. 200 BCE) Classical to early medieval periods (c. 200 BCE – c. 1200 CE) Late medieval period (c. 1200 – 1526 CE) Early modern period (c. 1526–1858 CE) Modern period and independence (after c. 1850 CE) Historiography See also References Further reading External links Navigation menureviewedee[update]The crisisThe Evolution and History of Human Populations in South Asia: Inter-disciplinary Studies in Archaeology, Biological Anthropology, Linguistics and GeneticsThe Ancient Indus: Urbanism, Economy, and Society"Indus River Valley Civilizations"The Ancient Indus: Urbanism, Economy, and Society"India before the British: The Mughal Empire and its Rivals, 1526–1857"Why Europe Grew Rich and Asia Did Not: Global Economic Divergence, 1600–1850Development Centre Studies The World Economy Historical Statistics: Historical StatisticsDeveloping cultures: case studiesEthnic Groups of South Asia and the Pacific: An Encyclopedia: An Encyclopedia"Indian Economy During British Rule""Economic Impact of the British Rule in India | Indian History"The Bone Readers: Science and Politics in Human Origins Research"Out of Africa: new hypotheses and evidence for the dispersal of Homo sapiens along the Indian Ocean rim"10.3109/0301446100363924920334598"Genetic and archaeological perspectives on the initial modern human colonization of southern Asia"2013PNAS..11010699M10.1073/pnas.1306043110369678523754394"Edakkal Caves|Places Around in Wayanad"Protecting megaliths to keep history alive The Hindu daily"Archaeologists rock solid behind Edakkal Cave"The Global Prehistory of Human Migration"Indus Valley 2,000 years older than thought"the original"Stepwells -Cosmology of Subterranean Architecture as seen in Adalaj"A History of Ancient and Early medieval India : from the Stone Age to the 12th century"Stone celts in Harappa"the original"Peoples and languages in pre-islamic Indus valley"the original"The Sindhi language"the originalThe Aryan chromosome"Fluvial landscapes of the Harappan Civilization"2012PNAS..109E1688G10.1073/pnas.1112743109338705422645375"Is River Ghaggar, Saraswati? Geochemical Constraints""An Ancient Civilization, Upended by Climate Change""Huge Ancient Civilization's Collapse Explained"2003GeoRL..30.1425S10.1029/2002GL0168222006QSRv...25.1283M10.1016/j.quascirev.2005.10.0122011QuInt.229..140M10.1016/j.quaint.2009.11.012Climate Change and the Course of Global History: A Rough JourneyA History of Ancient and Early Medieval India: From the Stone Age to the 12th CenturyA History of Ancient and Mediaeval India: From the Stone Age to the 12th CenturyAntennae SwordA History of IndiaA History of IndiaAn Introduction to Hinduism"India: The Late 2nd Millennium and the Reemergence of Urbanism"The Coinage of Ancient IndiaA Sanskrit reader: with vocabulary and notesPedigree: the origins of words from nature"Early Sanskritization. Origins and Development of the Kuru State"10.11588/ejvs.1995.4.823The Sanskrit epics, Part 2The City in South AsiaThe UpanishadsAn Introduction to HinduismReligions of the World, Second Edition: A Comprehensive Encyclopedia of Beliefs and PracticesA History of Ancient and Early Medieval India: From the Stone Age to the 12th CenturyEarly India: From the Origins to AD 1300Republics in ancient Indiapp. 83ff"Magadha Empire""Lumbini Development Trust: Restoring the Lumbini Garden"the originalThe great armies of antiquityarchived"The Achaemenid Persian Empire (550–330 B.C.)""East–West Orientation of Historical Empires"1076-156X"Dinner on the Grand Trunk Road"10.2307/32502263250226"Silappathikaram Tamil Literature"the originalManimekalai – English transliteration of Tamil originalIndian Temple Architecture: Form and Transformation : the Karṇāṭa Drāviḍa Tradition, 7th to 13th CenturiesBuddhist ArchitectureA History of India"The World Economy (GDP) : Historical Statistics by Professor Angus Maddison"The World Economy – Volume 1: A Millennial Perspective and Volume 2: Historical Statistics10.2307/32502140004-36483250214A Comprehensive History of India: Volume 2Between the Empires: Society in India, 300 to 400Emergence of Viṣṇu and Śiva Images in India: Numismatic and Sculptural Evidence"Parthian Pair of Earrings"the originalThe Medical Times and Gazette, Volume 1Greatest emporium in the worldThe Cambridge History of Ancient China: From the Origins of Civilization to 221 BCBuddhist Records of the Western WorldArchaeology in Soviet Central AsiaThe Grandeur of Gandhara: The Ancient Buddhist Civilization of the Swat, Peshawar, Kabul and Indus ValleysIndian Sculpture: Circa 500 B.C.-A.D. 700"The History of Pakistan: The Kushans"Gupta Dynasty – MSN Encartathe original"India – Historical Setting – The Classical Age – Gupta and Harsha""Gupta Dynasty, Golden Age Of India"the original"The Age of the Guptas and After"the originalNumber Theory and Its History"Gupta dynasty (Indian dynasty)""Gupta dynasty: empire in 4th century"the original"The Story of India – Photo Gallery"The ASI say499315420"Pallava script"p. 145"CNG: eAuction 329. INDIA, Post-Gupta (Ganges Valley). Vardhanas of Thanesar and Kanauj. Harshavardhana. Circa AD 606–647. AR Drachm (13mm, 2.28 g, 1h)""Harsha""Sthanvishvara (historical region, India)""Harsha (Indian emperor)"Shyama Kumar Chattopadhyaya (2000) The Philosophy of Sankar's Advaita VedantaShankara's IntroductionShankara's Introduction19373677Shankara's IntroductionIs The Buddhist 'No-Self' Doctrine Compatible With Pursuing Nirvana?The Seven Spiritual Laws Of YogaIndia: The Ancient Past. A History of the Indian-Subcontinent from 7000 BC to AD 1200The Kashmir Series: Glimpses of Kashmiri Culture – Vivekananda Kendra, Kanyakumari (p. 57).Al-Hind: Early Medieval India and the Expansion of Islam, 7th–11th CenturiesHistory of GopāchalaLand of Two Rivers: A History of Bengal from the Mahabharata to MujibEuropean Trade and Colonial ConquestA History of IndiaA Comprehensive History Of Ancient India (3 Vol. Set)"The Last Years of Cholas: The decline and fall of a dynasty"the originalFascinating Hindutva: Saffron Politics and Dalit MobilisationGazetteer of the province of OudhAl- Hind: The slave kings and the Islamic conquest. 2"Shahi Family"The Cambridge history of Islam"Ameer Nasir-ood-deen Subooktugeen"Gazetteer of the Attock District, 1930, Part 1Land of seven rivers: History of India's GeographyTemple Desecration and Indo-Muslim StatesIslam in South Asia: A Short HistoryBeyond Orientalism: The Work of Wilhelm Halbfass and Its Impact on Indian and Cross-cultural StudiesThe Making of Terrorism in Pakistan: Historical and Social Roots of ExtremismOrnament in Indian ArchitectureA historical review of Hindu India: 300 B.C. to 1200 A.D."Indian States and Union Territories"Islam in South Asia: A Short HistoryA Brief History of the Indian PeoplesThe Modern ReviewDelhi Sultanate"Battuta's Travels: Delhi, capital of Muslim India"the original"Timur – conquest of India"the originalIndia HandbookBhaktiThe Four Denomination of Hinduism10.1007/s11407-008-9049-925691067"Vijayanagara Research Project::Elephant Stables"10.2307/26465262646526Historical Dictionary of the TamilsBihar General Knowledge DigestMapping Bihar: From Medieval to Modern TimesPopular Literature and Pre-modern Societies in South AsiaA manual of the Kistna district in the presidency of MadrasAncient Indian History and CivilizationFragmented Memories: Struggling to be Tai-Ahom in India"The Islamic World to 1600: Rise of the Great Islamic Empires (The Mughal Empire)"the originalDynasties: A Global History of Power, 1300–1800, p. 105"Whose fort is it anyway"10.1111/0020-8833.000532600793Development Centre Studies The World Economy Historical Statistics: Historical Statistics"India's Deindustrialization in the 18th and 19th Centuries"The Mughal Empire, p. 190"The Long Globalization and Textile Producers in India"The Mughal World: Life in India's Last Golden AgeAurangzeb: The Life and Legacy of India's Most Controversial KingIn the Shadow of the Taj: A Portrait of Agra"Iran in the Age of the Raj"p. 8610.2307/20539802053980Delhi, the Capital of IndiaAn Advanced History of Modern India"Journal of the Tanjore Maharaja Serfoji's Sarasvati Mahal Library"The Rediscovery of India: A New SubcontinentIslamic Renaissance In South Asia (1707–1867) : The Role Of Shah Waliallah & His SuccessorsAn Advanced History of Modern IndiaThe Great Maratha Mahadaji Scindia"Full text of "Selections from the papers of Lord Metcalfe; late governor-general of India, governor of Jamaica, and governor-general of Canada""The Discovery Of IndiaThe Sacred City of the Hindus: An Account of Benares in Ancient and Modern TimesResurrecting Banaras: Urban Space, Architecture and Religious BoundariesFaith & Philosophy of Sikhism"Missiles mainstay of Pak's N-arsenal"History Modern India By S.N. Sen"Sirajuddaula"ArchivedLongman History & Civics (Dual Government in Bengal)Madhya Pradesh National Means-Cum-Merit Scholarship Exam (Warren Hasting's system of Dual Government)A Military History of Britain: from 1775 to the PresentIndian Cultural Heritage Perspective For TourismHindu Rulers, Muslim Subjects: Islam, Rights, and the History of KashmirIndian HistoryAn Atlas and Survey of South Asian HistoryAn Historical Account of the British Trade Over the Caspian SeaIndian Merchants and Eurasian Trade, 1600–1750The Indian diaspora in Central Asia and its trade, 1550–1900From Constantinople to the home of Omar Khayyam: travels in Transcaucasia and northern Persia for historic and literary researchA journey from Bengal to England: through the northern part of India, Kashmire, Afghanistan, and Persia, and into Russia, by the Caspian-SeaA Second Journey through Persia, Armenia, and Asia Minor, to Constantinople, between the Years 1810 and 1816Reports from the consuls of the United States, 1887Portugal and its Empire, 1250–1800 (Collected Essays in Memory of Glenn J. Ames).: Portuguese Studies Review, Vol. 17, No. 1The Dutch Power in Kerala, 1729–1758http://mod.nic.inArchivedDossier Goa – A Recusa do Sacrifício InútilAn Imperial Crisis in British India: The Manipur Uprising of 1891The Truth of Babri Mosque"Kolkata (Calcutta) : History"the original"Robert Clive, Baron Clive, 'Clive of India', 1725–1774""The Transformation from a Pre-Colonial to a Colonial Order: The Case of India"10.2307/25955872595587A versatile geniusArchived10.1109/MWSYM.1997.602854"Rabindranath Tagore on Education"the original"Essay on 'Derozio and the Young Bengal Movement'"Poverty and Famines: An Essay on Entitlement and Deprivation"Plague"the originalPopulation Growth and Land Use"Reintegrating India with the World Economy""Census Of India 1931"A history of modern India, 1480–1950"'India's well-timed diversification of army helped democracy' | Business Standard News"Bal Gangadhar Tilak: Struggle for Swaraj"Participants from the Indian subcontinent in the First World War""Commonwealth War Graves Commission Annual Report 2007–2008 Online"the original1462689197110.1017/s0010417500016534178920Eurocentrism: a marxian critical realist critique"Ranjit Guha, "On Some Aspects of Historiography of Colonial India""10.2307/2168385216838510.7202/016593ar"Harvard scholar says the idea of India dates to a much earlier time than the British or the Mughals""In The Footsteps of Pilgrims""India's spiritual landscape: The heavens and the earth""India: A Sacred Geography by Diana L Eck – review"Modern India: The Origins of an Asian Democracy10.2307/21694222169422964322464"The Indian Subcontinent and 'Out of Africa 1'"254043308Encyclopedia of World ReligionsThe Evolution and History of Human Populations in South Asia: Inter-disciplinary Studies in Archaeology, Biological Anthropology, Linguistics and Genetics"The Early Paleolithic of the Indian Subcontinent: Hominin Colonization, Dispersals and Occupation History"Ancient Indian History and CivilizationAncient Indian Social History: Some Interpretationsthe originalIndia Before EuropeA Concise History of Modern India"The beginning of the historical period, c. 500–150 BCE"full textA History of Indiathe originalexcerpt and text searchexcerptexcerptAn Economic History of India: From Pre-Colonial Times to 1991excerpt and text searchexcerpt and text searchIndia as known to the ancient world10.1111/j.1468-0289.1985.tb00391.x2597191onlineThe History of India, as told by its own historians. The Muhammadan Periodonline editionHans William Brown research collection on 19th-century missionary work in India, 1882–1932, Ms. Coll. 1033, Kislak Center for Special Collections, Rare Books and Manuscripts, University of Pennsylvaniaee