*Thanks for relpy . Please find my answers below inline.*
On Fri, Oct 26, 2018 at 2:41 PM Kamuela Lau <kamuela....@gmail.com> wrote: > Hi, > > Just to confirm, are you asking about the following? > > For a particular query, you have a list of documents, and for each > document, you have data > on the number of times the document was clicked on, added to a cart, and > ordered, and you > would like to use this data for features. Is this correct? > *[ME] :Yes* > If this is the case, are you indexing that data? > *[ME]* : *Yes we are planing to index the data but my question is how we should store it in to solr .* * should i create dynamic field to store the click, cart and order data per query for document?.* * Please guide me how we should store. * > > I believe that the features which can be used for the LTR module is > information that is either indexed, > or indexed information which has been manipulated through the use of > function queries. > > https://lucene.apache.org/solr/guide/7_5/learning-to-rank.html > > It seems to me that you would have to frequently index the click data, if > you need to refresh the data frequently > * [ME] : we are planing to refresh this data weekly.* > > On Fri, Oct 26, 2018 at 4:24 PM Midas A <test.mi...@gmail.com> wrote: > > > Hi All, > > > > I am new in implementing solr LTR . so facing few challenges > > Broadly we have 3 kind of features > > a) Based on query > > b) based on document > > *c) Based on query-document from click ,cart and order from tracker > data.* > > > > So my question here is how to store c) type of features > > - Old queries and corresponding clicks ((query-clicks) > > - Old query -cart addition and > > - Old query -order data > > into solr to run LTR model > > and secoundly how to build features for query-clicks, query-cart and > > query-orders because we need to refresh this data frequently. > > > > What approch should i follow . > > > > Hope i am able to explain my problem. > > >
