Share this article:


  • Join our comunity:

Using the Scopus-SciVal Topics Flow to Research Machine Learning

By: , Posted on: January 28, 2019

We previously looked at how integrating Scopus and SciVal together into your research process creates a seamless discovery flow that enables you to delve further into specific topics and research trends. With the addition of SciVal’s Topic Prominence of Science to Scopus, you gain a much clearer picture of your overall research performance, as well as the most promising areas for future research. The higher the Prominence percentile, the greater the momentum of that particular Topic and consequently the more likely it is to attract funding. In this post, you’ll see exactly how an engineer, interested in artificial intelligence, would broaden her research focus using Scopus and SciVal.

The Search Begins Broadly, Then Narrows

Our artificial intelligence researcher has decided to pursue a research project on machine learning. Already knowing that the subject has connections to many other fields, she begins her search for recent articles on the primary Topic.

First, the researcher conducts a general document search in Scopus on “machine learning,” to cast a wide net that will uncover all related research.

As the screen shot indicates, she receives more than 117,000 results from various journals and subject areas, both open access and paid-for, from different sources and consisting of different document types.

With research about machine learning progressing quickly, the researcher realizes that content published more than a few years ago may not be relevant today. So she further narrows her search – to open access articles and conference papers published during the past five years, in the fields of engineering, mathematics and machine learning. This reduces the results to 3,970, a more reasonable amount for the researcher to evaluate for the most relevant or frequently cited articles.

The researcher then chooses an article on “Scalable nearest neighbour algorithms for high dimensional data” that seems relevant to her research. Scopus enables her quickly to get up to speed on the contents of the article by providing an abstract.

Below the abstract, the researcher can see the Topic with which the article is affiliated, namely “Binary codes | Image retrieval | ANN search.” The Prominence percentile shows the current momentum of the Topic. It is calculated by weighing three metrics for papers clustered in a Topic: recent citation counts, recent Scopus views and the latest CiteScores of the articles’ journals. Each article contributes to only one Topic and this particular Topic is in the 99th percentile globally. Seeing this result, the researcher now might want to find out more about the Topic.

A New Window on the Way to SciVal

With one click, the researcher opens a new window within Scopus, where she sees further information related to the Topic such as Representative documents.

These documents are very strongly linked within the Topic, and are intended to give the researcher a feel for the central research question of a Topic. Representative documents typically have many within-Topic links and a large fraction of their links within the Topic, and are highly cited for their age as well. The full list of representative publications can be found in the Trends module as part of the Topic character functionality.

In this window, the researcher also can see the top authors who have contributed the most publications to this particular Topic. This information could be valuable to her in finding potential future collaborators. In addition, she can analyze the top 50 keywords related to the Topic in a word cloud. The larger the size of the keyphrase in the cloud, the more relevant it is to the Topic. The phrases generated in green are those that grew in scholarly output from 2013 through 2017. The researcher also has the option to further analyze the Topic within the window by toggling to a table view that shows the relevance of each keyphrase and its percentage of growth.

Moving to SciVal Brings Trends into Focus

Assuming she is logged into both Scopus and SciVal, the Researcher can now click on the “Analyze in SciVal” button to access the Trends module. Here, by hovering over the graph, she can see the number of articles published on the Topic, and how that total has grown steadily over time. The researcher also notes that the Field-Weighted Citation Impact (FWCI) has been falling slightly over time…which may be worth examining further. However, the Topic is still a strong performer, receiving 80 percent more citations than the global average (as dominated by a FWCI of 1.80). SciVal also reveals that international collaboration on this Topic has increased from 30 to 92 articles.

This SciVal screen shows the same keyphrases word cloud available in the Scopus pop-up window, with the majority of green words showing how the scholarly output in this Topic is growing.

At this point, the researcher can toggle to see only the representative publications if she wishes. SciVal both lists and graphs the institutions, countries and regions most prolific on the Topic, and reveals the primary authors and journals, too.

The researcher decides she’s interested in learning more about the Top Authors which she can select via the inner-page navigation bar.

In the Authors section, the researcher can dig deeper into the most prolific authors related to the Topic. She can learn where these top researchers are located (institutions and countries), and if they are already known to her.

This further analysis may help her decide whether to pursue a collaboration with one or more of the Top Authors. Maybe the researcher will find that she’s already working with some of these authors; she could analyze this further in the Collaboration module.

What About Related Topics?

Our researcher has thoroughly explored her original Topic in both Scopus and SciVal. But what about similar Topics related to this area of research, which are also highly prominent? She can get some answers to that question by clicking on the “Related Topics” tab.

The researcher immediately sees that the Topics are ranked in several ways. In “relatedness” – a measure of the match of the keyphrases between one Topic and another during the past five years – she discovers “query processing,” “algorithms” and “similarity queries” at the top. The researcher can also identify how much research has been published on the Related Topics and the percentage of Prominence of each Topic. She might just uncover areas of research that previously weren’t on her radar, thus expanding her discovery.

The Social Sciences Benefit from SciVal’s Topic Prominence Too

Now that we’ve shown how an engineer can benefit from conducting research using the combined power of Scopus and SciVal, it’s time to outline a case study for a different type of researcher. In our next blog post, a researcher in “cultural sociology” will be analyzed, to reveal the Topics in which he is active…among other data. In the meantime, try out the one-two punch of Scopus and SciVal yourself.  And let us know what happens.


Connect with us on social media and stay up to date on new articles


Stay up-to-date on the latest news and announcements from Elsevier. Learn about new product launches, programs and additions to our extensive list of books, journals and publisher tools.