Computational Analysis and Deep Learning for Medical Care. Группа авторов

Чтение книги онлайн.

СКАЧАТЬ when the user just types laptop in the search rather than showing pages related to just laptops. The recommendation system is working on this principle in e-commerce sites.

      Stages in information retrieval:

      1 Finding documents

      2 Formulating queries

      3 Determining relevance

      4 Rank the Pages

      Types of search engines:

      1 Crawler-Based search engines

      2 Directories

      3 Hybrid search engines

      4 Meta-search engines

As of now, every second Google search engine receives 60,000 queries and 5 billion for a day [26]. From theses, nearly 3 billion queries are from location-oriented devices. With the increased usage of pervasive computing, the location of the device is crucial while suggesting a query [13]. Earlier, the search engines used the query log for suggesting the query. But, now, the location has become an important aspect of the query suggestion. The historical choices of the registered user’s content which is used for query in the search engines will be used for the clustering process [21, 22]. The different techniques used for search engines are shown in Table 2.1. If the related queries are not present in the query log, then employ the location-aware keyword query suggestion (LKS). Once the user prefers the location, the search engine has to optimize the query to that specific location. The LKS will provide the related information to the user based on the location and queried data.

Table 2.1 History of search engines.

Search Engines	Method	Year
	Query logs	2000–2010
	Query session data [3]	2010–2012
	Query topic models	2013
	Semantic relevance of the keyword query	2008–2012
	Location-aware keyword, user preferences	2012–till data

      Especially while looking for a query which has significance in location like Taj mahal, Charminar, etc., will be the iconic locations. In such cases, query suggestion framework must be given the weight to the specific location with the spatial distances of the retrieved documents.

2.3 Artificial Intelligence Perspective

      AI makes the machines learn automatically from their experiences [31]. In the query suggestions approach, the role of AI is crucial, as the increased usage of the mobile and handheld devices makes us work in this domain. The mobile devices are used for personalized purpose, it contains lot of user-related and meaningful data and their preferences. These devices will make the personalized data and the location of the user. Based on this data and preferences of the user, the AI component will suggest the query and recommends the user. The preferences may vary from time to time and location wise. It is crucial in analyzing and gathering data to prepare the query suggestion. The inclusion of the location may vary the preferences of the user. As the AI agent will know the location, it will suggest the query.

The advancement in the AI industry with the rapid developments of computing power makes it more powerful. Deep learning algorithms work more effectively based on the availability of huge data. These models make the system can prepare more accurate queries based on the sequences. Location-based queries become more important with the rapid usage of the mobile devices. Based on this, the right information is delivered to the right people, at the right time [29]. In most of the cases like the emergency responses also served based on the locations of the user [30].

      2.3.1 Keyword Query Suggestion

      Query suggestion enables the user to scrutinize a query with a single click; this became one of the most fundamental features of web search engines. In general, it is not clear when the user would turn to query suggestion; it depends on circumstances. In order to investigate when and how the user uses query suggestion, we analyzed three kinds of datasets obtained from a major commercial web search engine, comprising approximately 126 million unique queries, 876 million query suggestions, and 306 million action patterns of users. The analysis shows that query suggestion is often used.

      1 When the original query is a rare query [8].

      2 When the original query is a single-term query,

      3 When query suggestions are unambiguous,

      4 When query suggestions are generalizations or error corrections of the original query, and

      5 After the user has clicked on several URLs in the first search result page.

      The search engines are working to provide better query suggestion input, and that they should dynamically provide query suggestions according to the user’s current state. There are different types of approaches for keyword query suggestion. This can be classified into three categories: random walk–based, cluster-based, and learning to rank approaches. We briefly review the other methods from our observation; any of the given methods cannot consider the user location in query suggestion.

       2.3.1.1 Random Walk–Based Approaches

      This method uses graph structure for modeling the information that is provided by query log and then applies the random walk process on graph for query suggestion [6, 7, 10, 11].

       2.3.1.2 Cluster-Based Approaches

In this method, the query log is viewed as query URL bipartite graph [2]. By applying the clustering algorithm on vertices in the graph, query cluster can be identified. Then, user supplied query q and queries that belong to same cluster as q does not returned to the user as suggestion.

       2.3.1.3 Learning to Rank Approaches

      This approach is trained based on different type of query features like query performance prediction [14]. Given query q, a list of suggestion is produced based on their similarity to q in topic distribution space. Query recommendation is a core task for large industrial search engines. The query recommendations is mostly depends on the query similarity measures. These measures can be used for query expansion or query clustering.

      2.3.2 User Preference From Log

Based on the user queries, the refinement is done using three functions: these are shown in Table 2.2.

      1 Modification of the user query

      2 Expansion

СКАЧАТЬ