Search Article 
 
Advanced search 
Official publication of the American Biodontics Society and the Center for Research and Education in Technology
Home About us Editorial board Search Ahead of print Current issue Archives Submit article Instructions Subscribe Contacts Login 


 
 Table of Contents  
ORIGINAL RESEARCH
Year : 2017  |  Volume : 8  |  Issue : 1  |  Page : 3-7

A scientific forecast on dental research output within the next 20 years using exponential smoothing algorithm


1 Independent Research Scientist, Founder and Managing Editor of Dental Hypotheses, Isfahan, Iran
2 Department of Endodontics, Dental Research Center, School of Dentistry, Isfahan University of Medical Sciences, Isfahan, Iran

Date of Web Publication14-Mar-2017

Correspondence Address:
Jafar Kolahi
No 24, Faree 15, Pardis, Shahin Shahr, Isfahan, 83179-18981
Iran
Login to access the Email id

Source of Support: None, Conflict of Interest: None


DOI: 10.4103/denthyp.denthyp_2_17

Rights and Permissions
  Abstract 

Introduction: To report a scientific forecast of the number of published dental articles in the next 20 years. Materials and Methods: On October 12, 2016, to find all dental articles, PubMed was searched via the query “1800/1/1”[PDAT]: “2015/12/31”[PDAT] AND jsubsetd [text]. Relevant limitations were applied to find dental clinical trials, review articles, and free full-text dental articles. Consequently, all PubMed records were exported to a CSV file. To forecast the future dental research output using existing time-based data, the Exponential Triple Smoothing algorithm was used, which is an advanced machine learning algorithm. Data were analyzed by Microsoft Office Excel 2016. Results: Seventy-five (1940–2015) years of human attempts to publish dental articles were explored and 572490 records were found, from which 27244 (4.75%) articles were free full-text, 19238 (3.36%) were clinical trials, and 31853 (5.56%) were reviews. Researchers will publish 19195 dental articles in 2036, among which 917 (4.77%) articles will be clinical trials, 1474 (7.67%) will be review articles, and 5482 (28.55%) will be free full-text articles. Conclusion: Changes may be because of the quantity of research funds. The number of all types of dental articles will increase with an acceptable rate over the next 20 years. Of more interest, the number of free full-text articles will grow more rapidly than other article types.

Keywords: Dentistry, exponential smoothing, forecast, future, research


How to cite this article:
Kolahi J, Khazaei S. A scientific forecast on dental research output within the next 20 years using exponential smoothing algorithm. Dent Hypotheses 2017;8:3-7

How to cite this URL:
Kolahi J, Khazaei S. A scientific forecast on dental research output within the next 20 years using exponential smoothing algorithm. Dent Hypotheses [serial online] 2017 [cited 2017 Aug 22];8:3-7. Available from: http://www.dentalhypotheses.com/text.asp?2017/8/1/3/202023


  Introduction Top


Human beings have been interested in forecastingsince ancient times. With the scientific development of mankind, mathematical models have been used toforecast future events. The British creative thinker Sir Arthur Charles Clarke expressed three principles for forecasting, known as Clarke’s three laws:[1]

Clarke’s first law

When a distinguished but elderly scientist states that something is possible, he is almost certainly right. When he states that something is impossible, he is very probably wrong.

Clarke’s second law

The only way of discovering the limits of the possible is to venture a little way past them into the impossible.

Clarke’s third law

Any sufficiently advanced technology is indistinguishable from magic.

However, predictive time-series analytics include a variety of statistical methods, e.g., autoregressive moving average, exponential smoothing and artificial neural networks, as well as various sub-categories to analyze the contemporary and historical evidence to make forecasts about future events.[2],[3]

Nevertheless, the aim of this study is to provide a scientific forecast of the number of published dental articles in the next 20 years.


  Materials and Methods Top


On October 12, 2016, PubMed was searched using the following queries to find all types of dental articles, dental clinical trials, review articles, and free full-text dental articles:



  • “1800/1/1”[PDAT] : “2015/12/31”[PDAT] AND jsubsetd[text]


  • “1800/1/1”[PDAT] : “2015/12/31”[PDAT] AND jsubsetd[text] AND Clinical Trial[ptyp]


  • “1800/1/1”[PDAT] : “2015/12/31”[PDAT] AND jsubsetd[text] AND Review[ptyp]


  • “1800/1/1”[PDAT] : “2015/12/31”[PDAT] AND jsubsetd[text] AND “loattrfree full text”[sb]


Consequently, all PubMed records were exported to a CSV file. To forecast the future dental research output, Exponential Smoothing algorithm was used.[4],[5],[6] This is a very common method to generate a smoothed time series data, although current data are given relatively more weight in forecasting than older data. This method requires consistent intervals between its data points. The forecast predicts future values using existing time series data and the AAA version of the Exponential Triple Smoothing algorithm, an advanced machine learning algorithm.[4]

The exponential smoothing forecast is computed using the formula:[7]

New forecast = Old forecast + α (Latest Observation − Old Forecast), where α (alpha) is the smoothing constant.

Or more mathematically:

Ft = Ft-1 + α (At-1–F t-1)

where,

α = Smoothing constant.

Ft = Forecast for period t.

At = Actual value in period t.

As the audience of this article are healthcare providers and dental science researchers, mathematical equations and formulas are not discussed. More information about this method is available at https://en.wikipedia.org/wiki/Exponential_smoothing.

However, time series data were analyzed using polynomial trend-line analysis. All data analyses were carried out by Microsoft Office Excel-2016.[4]


  Results Top


Seventy-five (1940-2015) years of human attempts to publish dental articles were explored and 572490 records were found, from which 27244 (4.75%) articles were free full-text, 19238 (3.36%) were clinical trials, and 31853 (5.56%) were reviews. The three oldest articles were published in 1940 in a French journal entitled “L’ Information dentaire.” [Figure 1] shows the number of all types of dental articles and free full-text articles published each year from 1940 to 2015. A dramatic growth of the number of dental articles started from 1965. The histograms of all types of dental articles published each year and the number of dental clinical trials and review articles are shown in [Figure 2] and [Figure 3], respectively.
Figure 1: Number of all types of dental articles (Mean: 7817.098±4983.754 and Confidence Level (95%): 1179.635) and free full-text articles (Mean: 533.294±800.479 and Confidence Level (95%): 225.138) published and indexed in PubMed from 1940 to 2015. Also, the polynomial trend-line analysis of data is presented

Click here to view
Figure 2: Pareto chart plots and distribution of the number of all types of dental articles published each year in descending order of frequency. The orang line represents the total cumulative percentage. Box and whisker plots of number of all types of dental articles are shown on the left

Click here to view
Figure 3: Number of dental clinical trials (Mean: 369.653±326.903 and Confidence Level (95%): 91.010) and review articles (Mean: 597.226±442.342 and Confidence Level (95%): 121.924) published and indexed in PubMed from 1940 to 2015. Also, the polynomial trend-line analysis of data is presented

Click here to view


Researchers will publish 19195 dental articles in 2036; from which 917 (4.77%) articles will be clinical trials, 1474 (7.67%) will be review articles, and 5482 (28.55%) will be free full-text articles [Figure 4],[Figure 5],[Figure 6],[Figure 7].
Figure 4: Results of exponential smoothing algorithm for all article types (Forecast ±  95% C.I). Number of articles will increase from 14884 in 2015 to 19195 articles in 2036. Alpha parameter is 0.9 and beta parameter is zero, showing the forecast is based more on the recent data and trends. Gama parameter is zero, indicating the data and forecast are not influenced by the seasonal changes. Mean absolute scaled error (MASE) is 0.70 and symmetric mean absolute percentage error (SMAPE) is 0.03, indicating a high level of forecast accuracy

Click here to view
Figure 5: Results of exponential smoothing algorithm for clinical trials (Forecast ±  95% C.I). Number of articles will increase from 414 in 2015 to 917 articles in 2036. Alpha parameter is 0.75 and beta parameter is zero, showing the forecast is based more on the recent data and trends. Gama parameter is zero, indicating the data and forecast are not influenced by the seasonal changes. Mean absolute scaled error (MASE) is 3.26 and symmetric mean absolute percentage error (SMAPE) is 0.13, indicating an acceptable level of forecast accuracy

Click here to view
Figure 6: Results of exponential smoothing algorithm for review articles (Forecast ±  95% C.I). Number of articles will increase from 903 in 2015 to 1474 articles in 2036. Alpha parameter is 0.090 and beta parameter is 0.75, showing the forecast is based more on the recent data and trends. Gama parameter is zero, indicating the data and forecast are not influenced by the seasonal changes. Mean absolute scaled error (MASE) is 1.74 and symmetric mean absolute percentage error (SMAPE) is 0.10, indicating an acceptable level of forecast accuracy

Click here to view
Figure 7: Results of exponential smoothing algorithm for free full-text articles (Forecast ±  95% C.I). Number of articles will increase from 2286 in 2015 to 5482 articles in 2036. Alpha parameter is 0.13 and beta parameter is 0.13, showing the forecast is based more on the recent data and trends. Gama parameter is zero, indicating the data and forecast are not influenced by the seasonal changes. Mean absolute scaled error (MASE) is 6.87 and symmetric mean absolute percentage error (SMAPE) is 0.13, indicating an acceptable level of forecast accuracy

Click here to view



  Discussion Top


In this study, we analyzed the dental articles published and indexed in PubMed from 1940 to 2015. A dramatic growth in the number of dental articles started from 1965. On the other hand, a sudden and sharpdeclineoccurredin the number of published dental articles during 1990–1992. Furthermore, the number of dental clinical trials decreased sharply during 2012–2015. These fluctuations might be due to the changes in the quantity of research funds. However, more effort is needed to find the pros and cons of these events.

Nevertheless, the forecasting results are promising. The number of all types of dental articles, dental clinical trials, review articles, and free full-text articles will increase with an acceptable rate during the next 20 years. Of more interest, the number of free full-text articles will grow more rapidly than the others.

Opportunities for planning in the future will be provided through forecasting. Scientific forecast on dental researches will help publishers and journal editors to devise a scheme for future publishing. If we subcategorize the forecasting in each field, we will easily recognize which subject has received more attention and which field is in focus. In addition, the major advantage of forecasting is funding and financing, it will provide us with a vivid outlook for future planning.

In this study, we had access to annual data of PubMed. If monthly or seasonal data were available, the forecast would be more reliable. We used Exponential Smoothing algorithm for forecasting. Other well-known methods, e.g., heuristic algorithms, willbe useful to criticize our results.[8]

Financial support and sponsorship

Nil.

Conflicts of interest

The authors have editorial involvement with Dent Hypotheses. There are no conflicts of interest.

 
  References Top

1.
Clarke’s three laws − Wikipedia, the free encyclopedia. https://en.wikipedia.org/wiki/Clarke%27s_three_laws [Last accessed on 3 Jun 2016].  Back to cited text no. 1
    
2.
Time Series Analysis − Statistics Textbook. http://documents.software.dell.com/statistics/textbook/time-series-analysis [Last accessed on 29 Oct2016].  Back to cited text no. 2
    
3.
Time series − Wikipedia. https://en.wikipedia.org/wiki/Time_series.  Back to cited text no. 3
    
4.
5.
Ostertago AE, Ostertag O. Forecasting using simple exponential smoothing method. Acta Electrotech Inform 2012;12:62-6.  Back to cited text no. 5
    
6.
Kalehar PS. Time series Forecasting using Holt-Winters Exponential Smoothing. http://www.it.iitb.ac.in/∼praj/acads/seminar/04329008_ExponentialSmoothing.pdf [Last accessed on 29 Oct2016].  Back to cited text no. 6
    
7.
Mahendra Gor R. Industrial statistics and operational management: Forecasting techniques. http://nsdl.niscair.res.in/jspui/bitstream/123456789/829/1/CHAPTER-6 FORECASTING TECHNIQUES- Formatted.pdf [Last accessed on 29 Oct2016].  Back to cited text no. 7
    
8.
Kokash N. An introduction to heuristic algorithms. http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.105.8050&rep=rep1&type=pdf [Last accessed on 29 Oct2016].  Back to cited text no. 8
    


    Figures

  [Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7]



 

Top
 
 
  Search
 
Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
Access Statistics
Email Alert *
Add to My List *
* Registration required (free)

 
  In this article
Abstract
Introduction
Materials and Me...
Results
Discussion
References
Article Figures

 Article Access Statistics
    Viewed411    
    Printed15    
    Emailed0    
    PDF Downloaded75    
    Comments [Add]    

Recommend this journal