If the subject is published widely with opponent results then yes

Yes why not, but you should know the correct reason why accuracy is decreasing. This is also a research you are telling people what not to do. In case you found the reason of decreased accuracy then publish your article after removing those issues.

For a conference paper: Generally Yes

For a research paper submitted to a journal: It depends on the journal.

To my knowledge, IEEE Aceess accepts articles reporting new experimental or measurement techniques, including negative results. However, it is an open access journal, you need to pay publication fees if your paper is accepted.

https://ieeeaccess.ieee.org/

Dear Shantanu Kumar Rahut ,

Now, there is another possibility here. If you use deep learning, you should try to modify the hyper-parameters in your model to check whether the negative results come from the bias of your parameters. If the results still negative, you can write a paper to report that.

If you need a course about deep learning for NLP, check the course from stanford.

https://cs224d.stanford.edu/

Deep learning has been widely used by NLP researchers. The following article:

Recent Trends in Deep Learning Based Natural Language Processing (https://arxiv.org/pdf/1708.02709) discuss deep learning application in NLP. Therefore, I suggest you to repeat the experiment with different data and also use other evaluation techniques to validate your results.

Yes ! Conclusion of such research gives the causes and situations of negative results. Learning from failure is equally important output of research and this triggerrs to another research invitations.

But your writing skill and presentation of results shall be key factors for succeeding with your publication.

Good luck to your efforts.

In my opinion, yes. The article will be interesting. The results from the examinations if they went the same after several experiments confirm the specific research thesis.

Best wishes

Sure you can by this way you give indication for other researches in the same field research .

I think you can

Yes, reporting and discussion negative (not conventional) results will certainly be interesting to be published.

Yes, it may be.

Results are not negative or positive. They may not be departure from conventional pattern.

Try best to explain such deviations. It should be an interesting challenge.

Of course you can.

History show that very often we learn a lot of things about negative results or approaches. Moreover it prevents other scientists to avoid this way.

I have read once a short section from a great mathematician that the fact that negative approaches was not published slow down advances in a lot of fields, in a great amount.

So ....

Agree with Herve Poulard

Yes,the results do not necessarily have to be positive and the reason for the negative results must be explained.

Yes you must write research paper though result negative...

Dear Shantanu,

Yes, i think you can explaining why the goal is not possible in that way.

Happy research!

~AT

I think it must be possible. Negative results are inportant to science.

It is another question, if you get it published in a well known Journal.

However I got over new journals that have been founded only to publish negative results. You should look for such a specialized journal for your paper!

Good luck for your project...

Two examples of such negative results journals:

https://www.negative-results.org/ and

http://www.jnr-eeb.org/index.php/jnr

A similar question to this one is available under the RG thread

https://www.researchgate.net/post/Suggest_any_journals_that_publish_negative_results_and_its_impact

Here is an even longer list of journals for negative results, see

"Several journals publish negative results.

Here are a few of them:

Journal of Negative Results in BioMedicine www.jnrbm.com/‎

Journal of Negative Results www.jnr-eeb.org/‎

Journal of Pharmaceutical Negative Results www.pnrjournal.com/‎

Journal of Interesting Negative Resultsjinr.org/

The All Results Journals: Chemwww.arjournals.com/‎

NEGATIVE RESULTS IN CHEMISTRY www.negativeresultsjournals.com/.../negative-results-in-chemistry.html‎"

Re-shared from the other RG Thread...

Certainly, negative results should be reported and published with proper reasons.

Result can be positive or negative, one has to present the data with proper explanation, with proper methodology.

In scientific research, there are basically no bad or good results, there are only results confirming or refuting the research thesis. Of course, the conditions must be met for conducting objective and independent scientific research. If these conditions are met, the tests were carried out on a representative sample, the tests were repeated and the test results confirmed, then the obtained results, regardless of whether they confirm or refute the previously formulated research thesis, may be considered as the success of scientific research.

On the other hand, the methodology and specificity of scientific research and research works looks a little different if instead of verifying the previously formulated research thesis, e.g. a technical model, a prototype of a new device is being built, the design and construction of which uses innovative technologies, materials and construction solutions , logistics and other. If an innovative device built in a research laboratory, a new technology is created, a new type of material is invented, a new type of energy source, etc. is built according to a predetermined purpose, then the results of this type carried out in a modern laboratory equipped with research technology may end up being full or incomplete. the implementation of a previously defined plan, goal, project, and thus may end with more or less success.

There are situations during scientific research and research that the positive results of research and research are greater than those previously identified. These types of situations occur when laboratories with modern technologies have been built as part of high-budget research programs run and financed by government research agencies and / or large technology corporations. An example is the large high-budget research programs implemented during space exploration programs coordinated by government astronautics agencies. This situation occurred during the NASA program coordinated by NASA in the 1970s to send manned space missions to the Earth's moon. At that time, new types of materials and innovative technologies were created in the research and development laboratories set up by the government agency, which quickly found application in various industries besides astronautics. This type of situation may be repeated in the 2020s or 2030s, when the program of sending manned space missions to Mars, coordinated and co-financed by the US government.

Negative results can happen in many occasions, also in small and low budget Setups...

I co-edit a number of journals. From my perspective, we can still learn from negative results.

If the subject is published widely with opponent results then yes

I see no reseason not to publish negative results provided the trends are well explained.

Shantanu Kumar Rahut , The result that you obtained from your experiment must not be positive or negative. The important fact is you got a result and all you have to do is to identify why you have got that result. Find an explanation of that. You can draw comparison with other works and justify why you got that result. This can be considered as a good work and of course publishable.

Negatibve and possstive results are relative interpratations.

One can write a paper in case your results depart from a certain expected pattern.

My view is that as long as the research methodology is sound, then results should be published. Some journals will publish papers even if results are not significant.

Of course it is valuable result. If a phenomenon that you investigate have some strongly grounded behavior, then more data you collect, it is more convergent to the conclusion of its behavior. Since you are getting more blur situation as you collect more data, then you are loosing the focus if the expected behavior. This could also be expected, not an irregular situation. You can publish this too, but there is question:what is the aim of research, what is the value and focus? What benefits a reader can have? Maybe you change the type of data you collect, or change the expected behavior you target. Then you maybe get better results.