Good morning, Paul! Thank you for your comments.

Before 2011, I wrote many Japanese papers about OLDFs and book. But my several papers were rejected by noisy reasons after 2011. Last year, I realized my Japanese papers are never accessed from the world because researchers enter CiNii system in Japan and download papers in charge. Last Nov. I decided to apply the English papers that is offered by digital libraries. It is regretful for me not to upload ICORES2014 accepted paper. You can find the abstract on Internet World. Full text must be download from the HP as same as IEEE. Therefore, I upload the draft before accepted version. But ICORES is good international conference. You know IFORS, INFORMS and EuroOR request only abstract.  After 1996, I present OR international conferences, but no response because of abstract. 

Please advice from now.

I have some published papers on discriminant analysis which was  used on drug data and mode of delivery of an expectant mother

Dear Oluwafemi

I'm sorry not to respond your answer because I'm crazy about my research.

I read "A comparison of the performance of LDF and the logistic....(October 2014)".

It is very interested, but you had explain the results and discussion in detail  to delete the some part of explanation of logistic regression. If I discriminate your data, I look for the best model among all possible 31 models. If so, we need not discuss the individual statistics because the best model is accepted in total.

Dear Dr. Suichi Shinmura:

I work in the field of complex systems, speciffically in Engineering Systems. But medical systems are complex systems too, so I think the methods to be used are the same. In my work field there are included: decision making, systems analysis  data proccessing (that involve non linear and linear multipliple regresion analysis). The most of my work it is written in Spanish Language, but I have some publications in English too. I propose you to read the Mononography "Analysis and Synthesisof Engineering Systems, thinking about medical decisions making systems.This monography is available in my researchgat profile. If you think my work could be useful for yours, please to send me some concrete first proposals, and let begin the cooperation,

Best wishes,

Dr.Jose Arzola

Dear Jose

Thank you for your offer.

Please some times for me.

Until now, my "k-fold cross validation " offers us the 95% CI of error rate and coefficients. And I get marvelous results  of error rates. The best model with minimum mean of error rates tell us many useful knowledge. But I am not obtain the useful meaning  of the 95% CI of coefficient. Recently, I found the median of coefficient distribution become a trivial LDF in the pass/fail determination. I'm writing the paper. So, please some time, but I have no knowledge for the enginieering. 

Dear Jose

I read your mononography that deals wide research area. My research is only sub-module in your research. I'd like to establish the new world of the discriminant functions after Fisher. Probably, I will complete next year. If so, can you tell me my theory is useful  in your sub-module. In that chance, we negotiate.

Dear Oluwafemi

Can you choose the good discriminant model among 11 linear separable models of Swiss banknote data by logistic regression? Flury et al. collected this data for the evaluation of discriminant function. I found the 11 models incliding (X4,X6) are linear separable. I hope someone recommends the good model among 11 models by Fisher’s LDF, or logistic regression, or hard margin SVM (H-SVM) or soft margin SVM (S-SVM). My claim is you cannot choose proper model based on the existing statistics or test.

Excel data: Yi=1:genuine bills, -1:counterfeit bills

Dear Shuichi Shinmura:

I am ready to work with you,maybe helped by somebody of my collective. Please to send me a concrete proposal and some initial literature.

Best wishes, Arzola

Dear Jose. My proposal is very simple. If there are some data accepted by the article, I discriminated the data by seven LDFs and write a draft. Co-worker revised the draft to the article. The first author is decided by negotiation. Co-author must agree his or her original results may be wronger than the new one. If departments or the society offers data, we need the agreement. They make a agreements draft. Initial literatures are my update articles after 2014. However, I am writing new four papers now. I am afraid to open these papers beforehand. Can you tell me whether this action cause the problem or not with the conference or articles.

If Oluwafemi propose the co-authur, there are some points we must discuss.

* There is no mesurements by medical instruments. In Africa, I guess the height, weight, age, gender of mothere and infant are easy to collect. But, there is no explanation. He must explain he choose these variables.

* He must explain about the rule of the Ceaserian section or natural delivery. Who judge and choose the cases.

* He choose the model by the statistics such as Wald etc. But I doubt these statistics. I think the best model with minimum M2 solve the problem instead of their approach. 

* I am afraid the new paper will be accepted by the same journal.

Thank you so much for your help.