Tony Russell-Rose Hi Marina If you mean mining logs for the purposes of search analytics (i/e. insight into comsumer behaviour through their queries) then Lou Rosenfeld’s book is a pretty good start. It cover all the basic techniques from an IA and UX perspective, which should map onto your CEM angle. Not sure about the BI piece – what did you have in mind here?
Marina Santini Hi Tony, do refer to “Search Analytics for Your Site CONVERSATIONS WITH YOUR CUSTOMERS By Louis Rosenfeld. Rosenfeld Media, July 2011″ ? As for the BI bit, I was wondering whether query logs can give us hints of users’ “future” needs, i.e. needs that are not currently met but that can be fulfilled in the future, In a word, I would like to see whether query logs can be used for predictions and for the discovery of new business trends. Cheers, Marina
Tony Russell-Rose Yes, that’s the one. Re the prediction piece, it’s an interesting idea, but I’m not sure query logs are quite that forward looking. They may signal intent, but bounded by the session more than anything, and the ‘zeitgiest’ element of those I’ve looked at was very unclear/fragmented. Woudl social media not be a better place to start?
Jan Jasik • In SOA environment, one could collect data using SLA extended metrics as a part of ESM/WSM environment… have not seen anyone planning or mining query logs.
Marina Santini • I am very curious about the possibility of mining query logs for “meaningful” information.
Although I am little bit worried about the semantic fuzziness of users’ queries…
Jan Jasik • … BI systems designers and administrators use queries to evaluate its effectiveness and making improvements to “context-based routing” frequently used by UIs of those systems. The aggregates of ‘semantic fuzziness’ (point well taken) have a purpose at a critical mass level; individually, perhaps not so much. Unless you can find it useful by mapping to something: environment, event, behavior, … discovering e.g. intentions (by inference)?
Marina Santini • There is quite a lot of literature on the concept of “query intent”… May be I should follow that path and see how it can be applied to BI e CEM. This is what Yahoo! is doing with query intents : http://labs.yahoo.com/node/189
Marina
From LinkedIn: Enterprise Search Engine Professionals (Group)
Helge Legernes • Have a look at Splunk !
Marina Santini • Thanks Helge. I will!
Cheers, Marina
Marina Santini • I just found this interesting abstract: “Mining associations from web query logs” (http://labs.yahoo.com/node/189), and I was wondering whether it would be more convenient to focus on how “query intents” can be used for BI & CEM…
Marina
From LinkedIn: UTMA – Ubiquitous Text Mining and Analytics (Group)
Discussion link: http://lnkd.in/MpZ-bJ
Ina Lauth • Dear Marina,
please look first for the research work of Ricardo-Baeza-Yates, Director of Yahoo Research Lab in Barcelona (and his team). Here the latest research publications:
ECIR 2012: http://ecir2012.upf.edu/accepted.html
Ina Lauth • I also recommend the Tutorials and lectures on “Videolectures.net” : e.g. the Tutorial on Query Log Mining” by Ricardo Baeza-Yates and Fabrizio Silvestri..there are a couple of more lectures on this topic there…
I give you a few hints from our research community on this topic , but what I would do is to read through one or two of these papers that would come as near to your topic as possible and then contact the experts directly. Ask them your questions directly, they are always willing to help (this is my personal experience). You can reach them over the websites of their institutes. They all have huge experience with industrial and EU-projects projects on these topics.
From LinkedIn: KD2U – Knowledge Discovery in Distributed and Ubiquitous… (Group)
Discussion link: http://lnkd.in/777JrQ
Mark Sanderson • On a specific topic, there was this paper on mining the logs for spell correction
S. Cucerzan and E. Brill, BSpelling correction
as an iterative process that exploits the
collective knowledge of web users,[ in Proc.
EMNLP, 2004, pp. 293–300.
From LinkedIn
Tony Russell-Rose Hi Marina If you mean mining logs for the purposes of search analytics (i/e. insight into comsumer behaviour through their queries) then Lou Rosenfeld’s book is a pretty good start. It cover all the basic techniques from an IA and UX perspective, which should map onto your CEM angle. Not sure about the BI piece – what did you have in mind here?
Marina Santini Hi Tony, do refer to “Search Analytics for Your Site CONVERSATIONS WITH YOUR CUSTOMERS By Louis Rosenfeld. Rosenfeld Media, July 2011″ ? As for the BI bit, I was wondering whether query logs can give us hints of users’ “future” needs, i.e. needs that are not currently met but that can be fulfilled in the future, In a word, I would like to see whether query logs can be used for predictions and for the discovery of new business trends. Cheers, Marina
Tony Russell-Rose Yes, that’s the one. Re the prediction piece, it’s an interesting idea, but I’m not sure query logs are quite that forward looking. They may signal intent, but bounded by the session more than anything, and the ‘zeitgiest’ element of those I’ve looked at was very unclear/fragmented. Woudl social media not be a better place to start?
Marina, have a look at Sematext Search Analytics. It’s free: http://sematext.com/search-analytics/index.html
From: Enterprise Architecture: Tactical. Strategic. Visionary. (LinkedIn Group)
Jan Jasik • In SOA environment, one could collect data using SLA extended metrics as a part of ESM/WSM environment… have not seen anyone planning or mining query logs.
Marina Santini • I am very curious about the possibility of mining query logs for “meaningful” information.
Although I am little bit worried about the semantic fuzziness of users’ queries…
Jan Jasik • … BI systems designers and administrators use queries to evaluate its effectiveness and making improvements to “context-based routing” frequently used by UIs of those systems. The aggregates of ‘semantic fuzziness’ (point well taken) have a purpose at a critical mass level; individually, perhaps not so much. Unless you can find it useful by mapping to something: environment, event, behavior, … discovering e.g. intentions (by inference)?
Marina Santini • There is quite a lot of literature on the concept of “query intent”… May be I should follow that path and see how it can be applied to BI e CEM. This is what Yahoo! is doing with query intents : http://labs.yahoo.com/node/189
Marina
From LinkedIn: Enterprise Search Engine Professionals (Group)
Helge Legernes • Have a look at Splunk !
Marina Santini • Thanks Helge. I will!
Cheers, Marina
Marina Santini • I just found this interesting abstract: “Mining associations from web query logs” (http://labs.yahoo.com/node/189), and I was wondering whether it would be more convenient to focus on how “query intents” can be used for BI & CEM…
Marina
Otis Gospodnetić • Marina, have a look at Sematext Search Analytics. It’s free: http://sematext.com/search-analytics/index.html
Marina Santini • Thanks Othis!
From LinkedIn: UTMA – Ubiquitous Text Mining and Analytics (Group)
Discussion link: http://lnkd.in/MpZ-bJ
Ina Lauth • Dear Marina,
please look first for the research work of Ricardo-Baeza-Yates, Director of Yahoo Research Lab in Barcelona (and his team). Here the latest research publications:
ECIR 2012: http://ecir2012.upf.edu/accepted.html
Ina Lauth • I also recommend the Tutorials and lectures on “Videolectures.net” : e.g. the Tutorial on Query Log Mining” by Ricardo Baeza-Yates and Fabrizio Silvestri..there are a couple of more lectures on this topic there…
I give you a few hints from our research community on this topic , but what I would do is to read through one or two of these papers that would come as near to your topic as possible and then contact the experts directly. Ask them your questions directly, they are always willing to help (this is my personal experience). You can reach them over the websites of their institutes. They all have huge experience with industrial and EU-projects projects on these topics.
Ina Lauth • If you need query data:
Google BigQuery
is a scalable, easy to use, web service that lets you do interactive analysis of massive datasets-up to billions of rows. BigQuery is now available to the public:
http://www.kdnuggets.com/2012/05/google-bigquery-data-analytics.html?utm_source=twitterfeed&utm_medium=facebook&utm_campaign=tfb&utm_content=FaceBook&utm_term=analytics
Marina Santini • Great! Thanks a lot, Ina.
From LinkedIn: KD2U – Knowledge Discovery in Distributed and Ubiquitous… (Group)
Discussion link: http://lnkd.in/777JrQ
Mark Sanderson • On a specific topic, there was this paper on mining the logs for spell correction
S. Cucerzan and E. Brill, BSpelling correction
as an iterative process that exploits the
collective knowledge of web users,[ in Proc.
EMNLP, 2004, pp. 293–300.