Tag: TG_FRESH

The number of documents found in the Web-Base

The number of documents found in the database of the fast -butter

The number of documents found in Realtime

QuickDocCount * 100 / WebDocCountFresh

Peak size (in watches), 50 top documents are taken into account from each fast -butter machine

The total number of documents from the large base that have passed on Fastrank

The share of documents from a large base that have been created on Fastrank created in recent months

What percentage of fast documents came from the domain, which found the most results

What percentage of fast documents came from the second (according to the number of results) domain

How long have long been read from the disk database of fast -butter

Max(1.0 / (1 + difference between current date and date from url))

The number of words in the request

Is the request for navigation

The popularity of the request. The result of the sorcerer's work

The commercial request

The request mentions the commodity category

The request mentions vendor

In the request, the name of the organization (example: Gazprom, Gazprom)

The request is present in the current or next year

How much was the speed of this request before in the database of a fast -butter, the average in 2 weeks

How much was used to be in this request in Realtime, average in 2 weeks

Median news wizard shows rate for query doppelganger among last 6 30-days intervals

Number of news topics with entity from query doppelganger for last 6 months

Timestamp launching basic Fasttier

Number of documnets prepared for News Wizard

Is query foreign

Number of videos in News Wizard

Number of Yandex.News stories in News Wizard

1 if NewsWizard is a single story one, -1 else

Was there a request from several months to a couple of weeks ago

Component of the NW factor, compliance with the request of the headings of news documents from the I.Noos (according to N-grams)

Ratio between clicks on fresh results in last half hour and hour

Ratio between fresh surplus and total number of queries

Ratio between wins and wins + losses for fresh results

Ratio between clicks on news results in last half hour and hour

Ratio between news surplus and total number of queries

Ratio between wins and wins + losses for news results

Same as QuickClicksLastHourRatio, calculated using doppelgangers

Same as QuickSurplusToQueriesRatio, calculated using doppelgangers

Same as QuickWinsToTotalRatio, calculated using doppelgangers

Same as NewsClicksLastHourRatio, calculated using doppelgangers

Same as NewsSurplusToQueriesRatio, calculated using doppelgangers

Same as NewsWinsToTotalRatio, calculated using doppelgangers

Weighted rate of samohod in WEB L3 top-10

Rate of samohod in WEB L3 top-10

Rate of samohod in WEB L2

The number of fresh videos in the top 20 Y. Video

The ratio found in 3 days to the number of documents with non -expert values ​​found for all the time TRP1

Similarly, TRP1Histo3DFRACTION with non -Hasallwordstrfm value

Similarly, HasallwordStrfmhisto3dfuction, in the numerator - the number of documents over the past 5 hours

Similarly, TRP1HISTO3DFRACTION, but for Videop and Videoquick sources, Mediabestlrsyn factor

Similarly, Videomediabestlrsynhisto3dfuction, in the numerator - the number of documents over the past 5 hours

The linear combination of found in the last 5 hours, day and 3 days, divided into the total number of documents found for which PassagelegacyTR is not 0, only for Videop and Videoquick sources

Similarly TRP1HISTO3DFRACTION, but for sources of Videop and Videoquick

A metafator predicting whether the request will be fresh through n hours

There is a date in the request; It is considered to be 1 / (1 + the difference between the date of the query task and the date in the request in full days)

There are some news markers in the query

Ratio between number of found documents for 1 hour and number of found documents for 5 hours in QUICK source, such that the value of feature TxtPairEx is greater than 0.8

Ratio between number of found documents for 3 hours and number of found documents for 5 hours in QUICK source, such that the value of feature TxtPairEx is greater than 0.8

Ratio between number of found documents for 1 hour and number of found documents for all time in QUICK source, such that the value of feature TxtBreakEx is greater than 0.5

Ratio between number of found documents for 24 hours and number of found documents for all time in QUICK source, such that the value of feature TxtHeadEx is greater than 0.8

Ratio between number of found documents for 7 hours and number of found documents for all time in QUICK source, such that the value of feature TxtHeadEx is greater than 0.3

Ratio between number of found documents for 12 hours and number of found documents for all time in QUICK source, such that the value of feature TxtHeadEx is greater than 0.3

Ratio between number of found documents for 24 hours and number of found documents for all time in QUICK source, such that the value of feature TxtHeadEx is greater than 0.3

Ratio between number of found documents for 12 hours and number of found documents for all time in QUICK source, such that the value of feature TxtHeadEx is greater than 0.2

Max CTR for all words in request per month

Average CTR for all words in request per month

Average value of fresh hosts feature, based on clicks on fresh results, which was calculated for queries with Fresh.IP > 0.5 for each host with shows threshold 50

Average value of fresh hosts feature, based on fresh surplus

Accumulate value of fresh hosts feature, based on fresh surplus, which was calculated for each host with shows threshold 50

Minimal difference between request time and document time for top-5 documents from d:fresh grouping

Maximal difference between request time and document time for top-5 documents from d:fresh grouping

Average difference between request time and document time for top-5 documents from d:fresh grouping

Decayed counter for this query for last day, using doppelgangers instead of exact query

Similarly, TRP1HISTO3DFRACTION, but for video sources, Mediabestlrsyn factor

Similarly, Videomediabestlrsynhisto3dfuction, in the numerator - the number of documents over the past 5 hours

The linear combination of found in the last 5 hours, day and 3 days, divided into the total number of documents found for which PassagelegacyTR is not 0, only for video sources

Similarly TRP1HISTO3DFRACTION, but for a video source

Same as QuickClicksLastHourRatioDopp, collect logs for 5 minutes before each aggregation

Same as QuickSurplusToQueriesRatioDopp, collect logs for 5 minutes before each aggregation

Same as QuickWinsToTotalRatioDopp, collect logs for 5 minutes before each aggregation

Same as NewsClicksLastHourRatioDopp, collect logs for 5 minutes before each aggregation

Same as NewsSurplusToQueriesRatioDopp, collect logs for 5 minutes before each aggregation

Same as NewsWinsToTotalRatioDopp, collect logs for 5 minutes before each aggregation

Same as QuickClicksLastHourRatioDoppFixed, but use data only from mobile search

Same as QuickSurplusToQueriesRatioDoppFixed, but use data only from mobile search

Same as QuickWinsToTotalRatioDoppFixed, but use data only from mobile search

Aggregated sum of subqueries' fresh intent

Aggregated sum of subqueries' QDF (quick doc fraction)

Average fresh intent of subqueries

Average QDF of subqueries

Aggregated sum of subqueries' surplus (desktop)

Aggregated sum of subqueries' surplus (mobile)

Average surplus of desktop subqueries

Average surplus of mobile subqueries

Part of highlighted words in titles

Average value of fresh hosts feature, based on fresh surplus (calculated for top1)

Average value of fresh hosts feature, based on fresh surplus (calculated for hosts presenting in top1 and having shows threshold 50)

Maximal value of surplus for similar fresh serps (combinations of 3, 4 and 5 urls are checked on L2 stage)

Maximal value of surplus for similar fresh serps (combinations of 3 urls are checked on L2 stage)

Maximal value of surplus for similar fresh serps (combinations of 4 urls are checked on L2 stage)

Maximal value of surplus for similar fresh serps (combinations of 5 urls are checked on L2 stage)

Maximal value of surplus for similar fresh serps (combinations of 3, 4 and 5 urls are checked on L3 stage)

Maximal value of surplus for similar fresh serps (combinations of 3 urls are checked on L3 stage)

Maximal value of surplus for similar fresh serps (combinations of 4 urls are checked on L3 stage)

Maximal value of surplus for similar fresh serps (combinations of 5 urls are checked on L3 stage)

Max value for similarity of query and documents from fresh top5 (in bigrams mode of search dssm model)

Mean value for similarity of query and documents from fresh top5 (in bigrams mode of search dssm model)

Similarity of query and first document from fresh top5 (in bigrams mode of search dssm model)

Culinary request (Bayesov classifier by hosts)

Percent of requests, containing good hosts for this query during last 3 days

Metafator, prediction of WINS on other factors

Metafator, Losses prediction on other factors

Metafator, prediction of the total DwellTime for other factors

Replacement for FI_PAIRS_CONTRAST (desktop version)

Replacement for FI_PAIRS_CONTRAST (mobile version)

Replacement for FI_RT_CONTRAST_QUERY (desktop version)

Replacement for FI_RT_CONTRAST_QUERY (mobile version)

Min aggregate for doc factors NewsAgencyRating

Max aggregate for doc factors NewsAgencyRating

Avg aggregate for doc factors NewsAgencyRating