Today, I want to show you an upcoming feature of SPMF, which is a new tool called the Algorithm Explorer to explore the list of algorithms and tools offered in SPMF. It is actually a simple tool, but I think it can be useful, as there are many algorithms.
Note that this is a previewof the tool, as it will released in the next version of SPMF(2.59).
To open the new Algorithm Explorer, in the GUI of SPMF, we have to choose:
Then, this will open the new tool called Algorithm Explorer, where algorithms are shown as a tree on the left, classified by categories, and we can see information about the selected algorithm on the right:
In the above picture, we selected the AFEM-Rules algorithm. Thus, we can see the name of the algorithm, the category, the authors of the implementation, the link to the example page, the input and output types as well as the parameters.
Searching for similar algorithms
Update: I also added two buttons that allows to search for algorithms that are similar to a selected algorithm. More precisely, we can search for (1) algorithms with the same input, output and mandatory parameters, and (2) algorithms with the same input and output but that may not have the same parameters.
For example, if we select the RuleGrowth algorithm for sequential rule mining on the left and click on this button:
It will highlight all algorithms that have the same input and output types as RuleGrowth:
And if we instead click on this button:
It will highlight the algorithms that not only have the same input and output types as RuleGrowth but also the same mandatory parameters:
In this case, we can notice that TRuleGrowth is not included anymore because although it has the same input type and output type as RuleGrowth, it has an extra parameter that is the window length.
Let me show you one more example. Let’s say that I choose a high utility itemset mining algorithm like EFIM. Then, I can quickly find that many algorithms have the same input and output types and also the same mandatory parameters:
Conclusion
That is the current version of this tool. I will think about other potential improvements. If you have any suggestions, you may tell me in the comments below, either to add more functions or improve the user interface. I will try to take them into account, when I have time.
In this blog post, I will give a brief report about the IEEE ICDM 2022 (International Conference on Data mining), which I have attended. It started on November 28th 2022. I have attended ICDM as a workshop organizer this year and as a co-author.
What is ICDM?
ICDM is a good conference for data mining andmachine learning research. I have attended it for several years. You can see for example my reports about ICDM 2021 and ICDM 2020 on this blog. The proceedings are published by IEEE, and the conference was held offline with some online participants (for those who could not attend in person due to the pandemic).
Proceedings and acceptance rate
The proceedings were made available online to authors on the first day of the conference.
For the main proceedings, at ICDM 2022, 885 submissions were received from 54 countries. All papers were reviewed using a triple-blind process. From this, 85 regular papers and 89 short papers were accepted. Thus, the full paper acceptance rate is 9.77% and the overall acceptance rate is 20%.
For the workshop proceedings at ICDM 2022, there were 326 submissions in 15 workshops. And 135 papers (41%) were accepted.
Day 1: Workshops, but some problems
The first day was dedicated to workshops. There were many workshops on a variety of emerging topics. In particular, I co-organized the UDML workshop on utility-driven mining and learning, which is the 5th edition this year, and has been quite successful this year with about 20 submissions, 7 accepted papers, and over 20 participants.
However, on Day 1, I faced problems with the online platform that was adopted for the conference. It is Zoom Events, which can be viewed as a special version of Zoom designed to host events such as conferences with more functionalities than the basic Zoom platform. Zoom events provide a message board and a schedule of all events as well as the possibility to chat with other participants. The problems that I and other participants encountered are as follow.
First, our workshop was supposed to start at 10:00 AM but there was no “start” button to allow us to start our workshop. Thus, hours before, I sent several messages to the organizers on the platform and by e-mail, but we got no answer before the start time of our workshop. There were also a few other workshop organizers that had the same problem and were talking about how to solve the problem in the chat. As we did not receive any answer, for our workshop we decided to create a public Zoom link and send it to authors by e-mail to start the workshop on time. Then, at around 10:40, I received an e-mail telling me that I could start our workshop but we had started it already using our Zoom link, 40 minutes earlier…
Second, multiple authors from our workshop did not receive instructions about how to login to the conference before our workshop started. From 7 papers in our workshop, this affected 5 papers. So before the workshop, I receive multiple messages from authors about how to login. In one case, there was also an author who could not login because he used a different e-mail for his zoom account than for registering to the conference. And all these problems were not solved before our workshop start time. So this is another reason why we had to use a public Zoom link for our workshop.
From these problems, I think there are two lessons to learn: (1) It is better to do a rehearsal in advance with workshop organizers if a new system is used (like it was done for ICDM in Singapore in ICDM 2020, for example) to avoid issues, and (2) it would have been much simpler to just use Zoom public links instead of using Zoom events, because this latter creates problems for logging due to the restriction on e-mails who can log in.
Day 2 – Opening ceremonyand …
On the second day, I wanted to attend the opening ceremony and check the keynote talks as this is usually very interesting. However, I did not receive any Zoom Events link for accessing the second day of the conference in my e-mail. So it seems that only the first day of the conference is online. That is unfortunate but I can understand as it is supposed to be an in person conference. So that will be the end of this post about ICDM 2022.
Conclusion
This was a blog post about the ICDM 2022 conference. Hope it has been interesting. Will be looking forward to next year’s ICDM 2023 conference.
In this blog post, I will explain how to use the jpackage tool that is provided with Java to (1) package the JAR file of SPMF into an EXE file for Windows, and (2) to create an installer for SPMF.
Requirements
It is necessary to have:
A computer with Windows (as I will give instructions for this platform, but you may do something similar on other operating system)
A recent version of the Java JDK (at least version 14) so that you have the jpackage command available.
The JAR file of SPMF. You can download it here: spmf.jar or from the website of SPMF.
The ICO file of SPMF (if you want to make an application that has an icon): SPMF.ico
How tocreate an EXE application of SPMF for Windows
Step 1. On a windows computer, create two folders /input/ and /output/ on the desktop.
Step 2. Put the files spmf.jar and SPMF.ico in the folder /input/ that you have just created.
Step 3. Open the command line of Windows and execute this command:
Then, this will create an installer, which looks like this:
This is the installation process on Windows 11:
And those are the files after installation:
Customization and generating installers for other platforms
There are also many other options offered by jpackage, including generating packages for other platforms. For more information see the documentation of the jpackage command.
Conclusion
This was just a short blog post to show how to package SPMF.jar into a native application. I think the process of using jpackage is quite simple. In the past, I had used some other commercial tools to create EXE files for Java programs but the process was more complicated. I am thus happy to have found jpackage.
This is to announce that a new version of SPMF has been released on the 27th November 2022. This version has 7 new pattern mining algorithms:
the HUCI-Miner algorithm to mine closed high utility itemsets and generators at the same time (thanks to Jayakrushna Sahoo et al. for the original code )
the FHIM algorithm to mine all high utility itemsets (thanks to Jayakrushna Sahoo et al. for the original code)
the HGB algorithm to mine non redundant high utility association rules (thanks to Jayakrushna Sahoo et al. for the original code)
the HGB-all algorithm to derive all high utility association rules from the non redundant high utility association rules (thanks to Jayakrushna Sahoo et al. for the original code)
algorithms for mining sequential patterns with flexible constraints in a time-extended sequence database (eg. MOOC data)
the SPM-FC-L algorithm fi (Thanks to Wei Song et al. for the original code)
the SPM-FC-P algorithm (Thanks to Wei Song et al. for the original code)
Besides, it has several new features such as:
(1) An integrated text editor to open output file (to give an alternative to the system’s default text editor)
(2) Some improvements to the graphical user interface, such as shown below, such as colors to highlight algorithm categories and a window icon:
And some bugs have been fixed.
Besides a new MOOC.txt dataset of sequences of courses with timestamps has been added to the dataset page of SPMF.
Thanks again to all users and contributors to SPMF!
In this blog post, I will talk about the 11th International Conference on Model & Data Engineering (MEDI 2022), which I have attended this week. It was held from the 21st to 23rd November 2022, in Cairo, Egypt, and also online.
MEDI is a good conference for modelling, database, and related topics. MEDI is ranked C in the CORE 2021 ranking. It has been held in various countries over the years such as: Egypt, Estonia, France, Morocco, Spain, Cyprus, Italy and Portugal.
The local organization of the 2022 edition of MEDI is by the Nile University.
This year, I am glad to play the role of Program Committee (PC) Chair at this conference. Below, I give an overview of the event.
Opening ceremony
The opening ceremony was at 10:00 AM. The conference was introduced, and an overview of the program was given. Below, I present some slides from the opening ceremony with some more details.
Countries were MEDI was held:
Organizations. Here is an overview of the committees behind MEDI 2022:
Paper selection. About the program, this year 65 papers have been submitted to the main conference, and from that 18 were accepted for long presentation (28%) and 11 for short presentation (17%). The program committee consisted of 60 researchers from 23 countries, which provided 200 reviews, for an average of 3.5 reviews per paper.
Proceedings. Full presentation papers are published in a Springer LNCS volume, while short presentation papers are published together with workshop papers in a Springer CCIS volume.
Totally, there was 190 authors and submissions were made from 18 countries on five main topics: (1) Modelling, (2) Image processing and diagnosis, (3) Machine learning and optimization, (4) Natural language processing and (5) Database systems.
Workshop. A workshop called DETECT was co-organized with MEDI 2022, and 4 papers were accepted in that workshop (44%).
Special issues. Two special issues are organized for best papers of MEDI 2022
Keynote talks
There was three keynote talks.
The first keynote talk was by Prof. Vincent S. Tseng from National Yang Ming Chiao Tung University about “Broad and Deep Learning of Big Heterogeneous Health Data for Medical AI: Opportunities and Challenges“.
The second talk was “A service-based approach to drone service delivery in Skyway networks” by Prof. Athman Bouguettaya from The University of Sydney, Australia
The third talk was “Safety and security are key considerations in the design of critical systems” by Dr. Colin Snook from University of Southampton, United Kingdom.
Best paper awards of MEDI 2022
The best paper awards were announced during the closing ceremony. These papers were selected based on the peer reviews and also based on the presentations given at the MEDI conference.
Conclusion
This is all for this blog post! Hope this has been interesting. Looking forward to MEDI 2023! In about 1 week, I will talk to you about ICDM 2022.
In this blog post, I will talk about the 15th International Conference on Multi-disciplinary Trends in Artificial Intelligence (MIWAI 2022), which was held as a virtual event on November 18th, 2022. I have attended this conference to present two papers related to pattern mining.
MIWAI is a conference that has been held every year in the pacific/Asia region, for 15 years. In the past, MIWAI has been held in countries such as Vietnam, India, Thailand, China, Malaysia and Brunei. I have also attended MIWAI in Malaysia in 2019 (see my blog post about MIWAI 2019 here).
Proceedingsof MIWAI 2022
This year, the MIWAI conference received 42 papers from which 19 were accepted, which includes 14 full papers (acceptance rate of 33% for full papers) and 5 short papers (acceptance rate of 45% for full + short papers).
The proceedings of the conference are published by Springer in a book from the LNAI series. Hence, the papers are well-indexed in various publication databases such as DBLP, which is good.
Schedule of the conference
The conference was organized during a single day (November 18) and online using the Zoom platform. It started with an opening session, followed by a keynote talk by Prof. Rapeepan Pitakaso from Thailand about “Artificial Multiple Intelligence System (AMIS) and its Applications”. Then, there was paper presentations organized as parallel sessions during the rest of the day.
Opening
During the opening, the organizers talked about the program, the review process, the organization, etc. It was nice to see several people that I knew already from MIWAI 2019. The organizers are very friendly and professional.
It was announced that next year, MIWAI 2023 will be in Hyderabad, India. The submission deadline is planned for 10th January 2023.
Then, MIWAI 2024 will be in Chongqing, China.
Awards were also announced. I am please that the best paper award was given to “LCIM: Mining Low Cost High Utility Itemsets“, which is my paper. There was also another paper who received an award.
Keynote talk
There was a keynote talk “AMIS – Artificial Multiple Intelligence System: Theory and Application” by Prof. Rapeepan Pitakaso from Ubon Ratchathani Unviersity, Thailand. The talk was about a system called Artificial Multiple Intelligent System (AMIS), that aims to combine multiple types of intelligence, just like humans do (verbal, linguistic intelligences, etc.). The system is based on heuristic algorithms and also combines CNN (Convolutional Neural Networks) and other techniques. Here are some pictures and slides from this presentation.
Papers on pattern mining
As readers of this blog knows, I am interested in the field of pattern mining. At the conference, there was two papers about pattern mining (my papers):
5121
Philippe Fournier-Viger, M. Saqib Nawaz, Yulin He, Youxi Wu, Farid Nouioua and Unil Yun
A good thing about this conference is that the registration fee is cheap. For one paper, it costs 250$ USD and for two papers, I spent only 350$ USD. This is much cheaper than many other conferences published by Springer and also IEEE. For example, an alternative that I considered would have been to publish in some European conferences published by Springer such as DAWAK or DEXA but registering two papers for those conferences would have cost me over 1240 euros instead of 350$ USD! This is a big difference. There are also many IEEE top conferences that are very expensive and have an increasing price in recent years. For example, this year, IEEE ICDM has a registration price of 1300 USD, while the price 10 years ago was about only 500$ USD. Thus, I appreciate that MIWAI offers registration at a very reasonable price. I think it can allow researchers from all around the world to more easily publish their papers, especially when research funding is limited.
Conclusion
MIWAI 2022 was a good conference. It is not a very large conference but there are some interesting papers of good quality, and the conference is well-organized. I will try to attend again for MIWAI 2023. Later this month, I will talk to you about the MEDI 2022 and ICDM 2022 conference.
This is a short post about how to change the language of Tableau if the option is not available in the Help Menu.
Close Tableau
On Windows, open Regedit
Search for “Language code” to find the keys containing the languages options of Tableau.
Change the three following keys: “Language code”, “Repository language” and “SamplesLanguage” to your preferred language. For example “en_US” is for English and “zh_CN” is for Chinese.
5. Start Tableau again and the language will have been changed.
Today, I will talk about some upcoming feature for the next release of SPMF (2.58), which is a simple integrated text editor. This new release should happen in about 1 or 2 weeks (as I am very busy recently) and will contain some new algorithms. But also, it will contain an integrated text editor. This may seem very strange? Why a text editor? I will explain briefly the idea in this blog post to give an overview of this feature as I am testing it now. If you have comments, you may leave them in the comment section below
Why a text editor in SPMF?
Well, in previous version of SPMF, there was a few options to open the output files produced by the algorithms: using the system’s default text editor (e.g. Notepad on Windows) or using the Pattern Viewer tool of SPMF. Although using the system’s text editor can be good, I was thinking that having a customized text editor could be interesting and it is actually not difficult to implement. So, I will explain how it works below.
The SPMF text editor
In the next release of SPMF, there will be a new option to open the output file using the SPMF text editor instead of the default system text editor.
The SPMF text editor looks like this:
It has some useful features such as showing the line count and the line and column numbers in the status bar at the bottom. This bar can also be hidden.
Also a useful feature is to always highlight the current line (which is not done by NotePad, for example):
Besides, it is possible to activate a “night mode” from the menu:
And there is also a search bar that is very convenient for highlighting some keywords in a file, and works like the search bar in some Web browsers:
Also, there is of course the possibility of changing the font and the font size.
Besides, all the user-defined preferences are saved. So next time that you open SPMF, the same font, font size, night mode preference, window size and location, and other preferences of the SPMF text editor are kept.
The SPMF text editor can also display the size of the last file that is opened:
And here is an overview of the menus:
It has the basic important functions such as “Line wrap” and “Word wrap”.
Conclusion
Interesting? You will be able to try it in the next version of SPMF to be released soon.
For now, the features are quite simple because the aim is to provide an alternative to the system text editor specialized for opening the output files of SPMF. It is not designed to compete with a more complex text editor and to be used as a general-purpose text editor (although it could).
A limitation of the SPMF text editor is that text files are loaded in memory so it is restricted to opening files that are not too big.
Many other features could be added like highlighting keywords in the output file. So there is something to think about. Which features would be useful? You may let me know what you think. If some features are not too hard to implement and useful, I may add them. I will also do some more debugging before it can be released.
If you have any suggestions or ideas, you can let me know in the comment section or by e-mail at philfv AT qq.com
In this blog post, I will talk about the IEEE DSAA 2022 conference, which was held from the 13th to the 16th October 2022.
What is IEEE DSAA?
DSAA 2022 is the 9th edition of the IEEE International Conference on Data Science and Advanced Analytics (DSAA). DSAA is an international conference that has been held in many countries, and focuses on data mining, data science, big data, machine learning and relatedtopics.
DSAA is a relatively young conference compared to some top data mining and machine learning conferences, but DSAA has become more and more successful over the years.
Location
This year, the DSAA conference was planned to be held in the city of Shenzhen, China. But due to the COVID-19 pandemic, the conference was held in online mode (local organization by Shenzhen University), and using Zoom as online videoconferencing platform.
Proceedings of DSAA 2022
The proceedings of DSAA are published by IEEE.
Authors could submit a paper to the main track of DSAA. But besides that, the DSAA conference hosts several special sessions on emerging topics. It is interesting for authors that papers accepted in the special sessions are published as regular papers in the proceedings of DSAA. This year, I have been co-organizer of a special session called DSSBA (1st Special Session on Data Science for Social and Behavioral Analytics), that has been quite successful with 5 papers accepted (more on that later).
Another interesting aspect about DSAA is that there are two special issues, respectively organized in the International Journal on Data Science and Analytics (JDSA) journal and the Machine Learning Journal (MLJ). Authors could submit papers to these special issues and then present the articles at the conference.
Keynote speaker from the main conference
This year, there was a good line up of keynote speakers:
Conference opening
The conference opening was on the 13th October. Several interesting information were given about the DSAA conference. Here are some slides from the opening:
Country distribution for application track:
Country distribution for the research track:
The paper acceptance rate statistics:
The main topics of papers published in DSAA:
Some awards were given with some of them receiving a cash prize of 1000$ USD.
A keynote talk was given in this special session by Prof. Yun Sing Koh from the University of Auckland, New Zealand. She presented some of her latest research work related to machine learning to tackle environmental science challenges. In particular, she presented two recent research projects published in the Machine Learning journal and in AAAI 2022, which are about air quality index inference and about algal bloom monitoring, respectively. Below, I share a few slides from her talk:
For more details about these two research projects, you can see these two papers:
Olivier Graffeuille, Yun Sing Koh, Jörg Wicker, Moritz K. Lehmann: Semi-supervised Conditional Density Estimation with Wasserstein Laplacian Regularisation. AAAI 2022: 6746-6754
Ben Halstead, Yun Sing Koh, Patricia Riddle, Russel Pears, Mykola Pechenizkiy, Albert Bifet, Gustavo Olivares, Guy Coulson: Analyzing and repairing concept drift adaptation in data stream classification. Mach. Learn. 111(10): 3489-3523 (2022)
In the DSSBA special session, we also had four paper presentations, where the last two papers are about pattern mining.
SA-FGDEM: A Self-adaptive E-Learning Performance Prediction Model Wang, Liping; Ye, Mingtao; Zhang, Guodao; Sheng, Xin; Zhang, Jingran
Heterogeneous Drift Learning: Classification of Mix-Attribute Data with Concept Drifts Zhao, Lang; Zhang, Yiqun; Ji, Yuzhu; Zeng, An; Gu, Fangqing; Luo, Xiaopeng
Fast Mining RFM Patterns for Behavioral Analytics Wan, Shicheng; Deng, Jieyin; Chen, Jiahui; Gan, Wensheng; Yu, Philip S
There was also several interesting paper presentations at the conference and talks but due to my busy schedule, I was not able to attend many of them. An interesting paper that I saw related to periodic pattern mining is:
Discovering Periodicity in Locally Repeating Patterns Alfred Krzywicki (University of Adelaide); Ashesh Mahidadia (Rich Data Corporation); Michael Bain (University of New South Wales)
Conclusion
Overall, the conference has been very interesting and well-organized. I will try to participate again to IEEE DSAA, next year.