Creating a Feature to determine popularity
$begingroup$
I am Building a Recommendation System in which i have Multiple Category , I want to Know how Popular is my Product in each Categories. For that I am considering Probabilty as one factor. For e.g I have 3 Categories (C1,C2,C3) So i am Calculating
(No. Of Times Particular Item in Ci is Purchased)/Total No. Of Item Purchased from Ci
This Gives me Probability Of Each Item Within a Category. Apart From that I am also trying to considered
Total no. of times a Item has been Purchased/Total No. of times It has been Viewed
But i am not getting how to use Purchase/View Ratio With Probabilty. Apart From that what another factor can i conclude
recommender-system feature-engineering
$endgroup$
bumped to the homepage by Community♦ 2 mins ago
This question has answers that may be good or bad; the system has marked it active so that they can be reviewed.
add a comment |
$begingroup$
I am Building a Recommendation System in which i have Multiple Category , I want to Know how Popular is my Product in each Categories. For that I am considering Probabilty as one factor. For e.g I have 3 Categories (C1,C2,C3) So i am Calculating
(No. Of Times Particular Item in Ci is Purchased)/Total No. Of Item Purchased from Ci
This Gives me Probability Of Each Item Within a Category. Apart From that I am also trying to considered
Total no. of times a Item has been Purchased/Total No. of times It has been Viewed
But i am not getting how to use Purchase/View Ratio With Probabilty. Apart From that what another factor can i conclude
recommender-system feature-engineering
$endgroup$
bumped to the homepage by Community♦ 2 mins ago
This question has answers that may be good or bad; the system has marked it active so that they can be reviewed.
add a comment |
$begingroup$
I am Building a Recommendation System in which i have Multiple Category , I want to Know how Popular is my Product in each Categories. For that I am considering Probabilty as one factor. For e.g I have 3 Categories (C1,C2,C3) So i am Calculating
(No. Of Times Particular Item in Ci is Purchased)/Total No. Of Item Purchased from Ci
This Gives me Probability Of Each Item Within a Category. Apart From that I am also trying to considered
Total no. of times a Item has been Purchased/Total No. of times It has been Viewed
But i am not getting how to use Purchase/View Ratio With Probabilty. Apart From that what another factor can i conclude
recommender-system feature-engineering
$endgroup$
I am Building a Recommendation System in which i have Multiple Category , I want to Know how Popular is my Product in each Categories. For that I am considering Probabilty as one factor. For e.g I have 3 Categories (C1,C2,C3) So i am Calculating
(No. Of Times Particular Item in Ci is Purchased)/Total No. Of Item Purchased from Ci
This Gives me Probability Of Each Item Within a Category. Apart From that I am also trying to considered
Total no. of times a Item has been Purchased/Total No. of times It has been Viewed
But i am not getting how to use Purchase/View Ratio With Probabilty. Apart From that what another factor can i conclude
recommender-system feature-engineering
recommender-system feature-engineering
asked Dec 26 '18 at 6:27
DheerajkhannaDheerajkhanna
63
63
bumped to the homepage by Community♦ 2 mins ago
This question has answers that may be good or bad; the system has marked it active so that they can be reviewed.
bumped to the homepage by Community♦ 2 mins ago
This question has answers that may be good or bad; the system has marked it active so that they can be reviewed.
add a comment |
add a comment |
1 Answer
1
active
oldest
votes
$begingroup$
You need to put your shoes into business people and communicate with people from that department on regular basis, trying to understand what the business's needs are. In e-commerce your second metric is Conversion Rate and is defined as:
CR = Orders/Views*
Please note there that there are two variations of Views. One that comes from Product List Page (aka PLP) that is mostly know as Impression, or the one comes from Product Detail Page(aka PDP) is where the user actually lands on the product detail page. The latter is more common in the CR calculation.
Your first one i.e. the popularity, that somehow I liked, can be used. Only be careful that it is going to be affected by number of items available in that category.
In general, I find it useful to think about edge cases when coming up with new measures and to see whether the resultant is desirable. What do I mean by edge cases, let's take your popularity score, by your definition, for example. Imagine the following two scenarios:
- popularity item 1 = 1/2 (very limited items in this category where item 1 is)
- popularity item 2 = 1000/2000 (large number of items in this category where item 2 is)
What you see here looks quite normal by the original definition from mathematics perspective. However, what it means from recommendation/business perspective that you have similar scores from two items from two different categories. Here the business may jump in and say looks we have a lot of more assortments in the category where item 2 is and often items in this category are in the warehouse or have faster delivery etc. etc., then we prefer our products from there to be recommended first. This is the point you realize that while results seem reasonable, at least according to the popularity definition, falls short on certain aspects.
Other ideas: I noticed you are leaning towards using multiple scores. That is a great idea. Having multiple factors may catches these corners and address multiple business needs at the same time. Like having both CR and popularity. Or may think of each item's profit contribution to the business and take that into consideration. Imagine you have a score in place that gives equal chances of recommending an iPhone or Wireless Mouse, however iPhone contribute way more to the business (it is more profitable), then you may want to push for that factor too. You may also like to look at Impression I mentioned if it is available from your tracking. Item's rating, delivery, Add_To_Cart, Wish_List, or Time_Spent_on_Item are other things to think about. Check this post to get some other ideas how things are roughly done in Amazon. Hope these help.
$endgroup$
add a comment |
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
});
}
});
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f43147%2fcreating-a-feature-to-determine-popularity%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
$begingroup$
You need to put your shoes into business people and communicate with people from that department on regular basis, trying to understand what the business's needs are. In e-commerce your second metric is Conversion Rate and is defined as:
CR = Orders/Views*
Please note there that there are two variations of Views. One that comes from Product List Page (aka PLP) that is mostly know as Impression, or the one comes from Product Detail Page(aka PDP) is where the user actually lands on the product detail page. The latter is more common in the CR calculation.
Your first one i.e. the popularity, that somehow I liked, can be used. Only be careful that it is going to be affected by number of items available in that category.
In general, I find it useful to think about edge cases when coming up with new measures and to see whether the resultant is desirable. What do I mean by edge cases, let's take your popularity score, by your definition, for example. Imagine the following two scenarios:
- popularity item 1 = 1/2 (very limited items in this category where item 1 is)
- popularity item 2 = 1000/2000 (large number of items in this category where item 2 is)
What you see here looks quite normal by the original definition from mathematics perspective. However, what it means from recommendation/business perspective that you have similar scores from two items from two different categories. Here the business may jump in and say looks we have a lot of more assortments in the category where item 2 is and often items in this category are in the warehouse or have faster delivery etc. etc., then we prefer our products from there to be recommended first. This is the point you realize that while results seem reasonable, at least according to the popularity definition, falls short on certain aspects.
Other ideas: I noticed you are leaning towards using multiple scores. That is a great idea. Having multiple factors may catches these corners and address multiple business needs at the same time. Like having both CR and popularity. Or may think of each item's profit contribution to the business and take that into consideration. Imagine you have a score in place that gives equal chances of recommending an iPhone or Wireless Mouse, however iPhone contribute way more to the business (it is more profitable), then you may want to push for that factor too. You may also like to look at Impression I mentioned if it is available from your tracking. Item's rating, delivery, Add_To_Cart, Wish_List, or Time_Spent_on_Item are other things to think about. Check this post to get some other ideas how things are roughly done in Amazon. Hope these help.
$endgroup$
add a comment |
$begingroup$
You need to put your shoes into business people and communicate with people from that department on regular basis, trying to understand what the business's needs are. In e-commerce your second metric is Conversion Rate and is defined as:
CR = Orders/Views*
Please note there that there are two variations of Views. One that comes from Product List Page (aka PLP) that is mostly know as Impression, or the one comes from Product Detail Page(aka PDP) is where the user actually lands on the product detail page. The latter is more common in the CR calculation.
Your first one i.e. the popularity, that somehow I liked, can be used. Only be careful that it is going to be affected by number of items available in that category.
In general, I find it useful to think about edge cases when coming up with new measures and to see whether the resultant is desirable. What do I mean by edge cases, let's take your popularity score, by your definition, for example. Imagine the following two scenarios:
- popularity item 1 = 1/2 (very limited items in this category where item 1 is)
- popularity item 2 = 1000/2000 (large number of items in this category where item 2 is)
What you see here looks quite normal by the original definition from mathematics perspective. However, what it means from recommendation/business perspective that you have similar scores from two items from two different categories. Here the business may jump in and say looks we have a lot of more assortments in the category where item 2 is and often items in this category are in the warehouse or have faster delivery etc. etc., then we prefer our products from there to be recommended first. This is the point you realize that while results seem reasonable, at least according to the popularity definition, falls short on certain aspects.
Other ideas: I noticed you are leaning towards using multiple scores. That is a great idea. Having multiple factors may catches these corners and address multiple business needs at the same time. Like having both CR and popularity. Or may think of each item's profit contribution to the business and take that into consideration. Imagine you have a score in place that gives equal chances of recommending an iPhone or Wireless Mouse, however iPhone contribute way more to the business (it is more profitable), then you may want to push for that factor too. You may also like to look at Impression I mentioned if it is available from your tracking. Item's rating, delivery, Add_To_Cart, Wish_List, or Time_Spent_on_Item are other things to think about. Check this post to get some other ideas how things are roughly done in Amazon. Hope these help.
$endgroup$
add a comment |
$begingroup$
You need to put your shoes into business people and communicate with people from that department on regular basis, trying to understand what the business's needs are. In e-commerce your second metric is Conversion Rate and is defined as:
CR = Orders/Views*
Please note there that there are two variations of Views. One that comes from Product List Page (aka PLP) that is mostly know as Impression, or the one comes from Product Detail Page(aka PDP) is where the user actually lands on the product detail page. The latter is more common in the CR calculation.
Your first one i.e. the popularity, that somehow I liked, can be used. Only be careful that it is going to be affected by number of items available in that category.
In general, I find it useful to think about edge cases when coming up with new measures and to see whether the resultant is desirable. What do I mean by edge cases, let's take your popularity score, by your definition, for example. Imagine the following two scenarios:
- popularity item 1 = 1/2 (very limited items in this category where item 1 is)
- popularity item 2 = 1000/2000 (large number of items in this category where item 2 is)
What you see here looks quite normal by the original definition from mathematics perspective. However, what it means from recommendation/business perspective that you have similar scores from two items from two different categories. Here the business may jump in and say looks we have a lot of more assortments in the category where item 2 is and often items in this category are in the warehouse or have faster delivery etc. etc., then we prefer our products from there to be recommended first. This is the point you realize that while results seem reasonable, at least according to the popularity definition, falls short on certain aspects.
Other ideas: I noticed you are leaning towards using multiple scores. That is a great idea. Having multiple factors may catches these corners and address multiple business needs at the same time. Like having both CR and popularity. Or may think of each item's profit contribution to the business and take that into consideration. Imagine you have a score in place that gives equal chances of recommending an iPhone or Wireless Mouse, however iPhone contribute way more to the business (it is more profitable), then you may want to push for that factor too. You may also like to look at Impression I mentioned if it is available from your tracking. Item's rating, delivery, Add_To_Cart, Wish_List, or Time_Spent_on_Item are other things to think about. Check this post to get some other ideas how things are roughly done in Amazon. Hope these help.
$endgroup$
You need to put your shoes into business people and communicate with people from that department on regular basis, trying to understand what the business's needs are. In e-commerce your second metric is Conversion Rate and is defined as:
CR = Orders/Views*
Please note there that there are two variations of Views. One that comes from Product List Page (aka PLP) that is mostly know as Impression, or the one comes from Product Detail Page(aka PDP) is where the user actually lands on the product detail page. The latter is more common in the CR calculation.
Your first one i.e. the popularity, that somehow I liked, can be used. Only be careful that it is going to be affected by number of items available in that category.
In general, I find it useful to think about edge cases when coming up with new measures and to see whether the resultant is desirable. What do I mean by edge cases, let's take your popularity score, by your definition, for example. Imagine the following two scenarios:
- popularity item 1 = 1/2 (very limited items in this category where item 1 is)
- popularity item 2 = 1000/2000 (large number of items in this category where item 2 is)
What you see here looks quite normal by the original definition from mathematics perspective. However, what it means from recommendation/business perspective that you have similar scores from two items from two different categories. Here the business may jump in and say looks we have a lot of more assortments in the category where item 2 is and often items in this category are in the warehouse or have faster delivery etc. etc., then we prefer our products from there to be recommended first. This is the point you realize that while results seem reasonable, at least according to the popularity definition, falls short on certain aspects.
Other ideas: I noticed you are leaning towards using multiple scores. That is a great idea. Having multiple factors may catches these corners and address multiple business needs at the same time. Like having both CR and popularity. Or may think of each item's profit contribution to the business and take that into consideration. Imagine you have a score in place that gives equal chances of recommending an iPhone or Wireless Mouse, however iPhone contribute way more to the business (it is more profitable), then you may want to push for that factor too. You may also like to look at Impression I mentioned if it is available from your tracking. Item's rating, delivery, Add_To_Cart, Wish_List, or Time_Spent_on_Item are other things to think about. Check this post to get some other ideas how things are roughly done in Amazon. Hope these help.
answered Dec 26 '18 at 8:46
Majid MortazaviMajid Mortazavi
1,6901220
1,6901220
add a comment |
add a comment |
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.
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fdatascience.stackexchange.com%2fquestions%2f43147%2fcreating-a-feature-to-determine-popularity%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
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