Several teachers in the industry have also discussed this issue. There is a teacher that data mining is just a laboratory. For this view he believes that the tools of data mining are also specialized, and the average person is difficult to apply, and the current data mining tool needs to be relatively deep. It is also like the development of the DOS era in the development of the operating system, and the XP system that reaches the popularization needs a process. Some teachers believe that data mining is now a well-ventured technology, because data mining itself is a marginal discipline, rarely has its own unique technical point. And its main related technologies are more difficult, statistics, etc. For tool development research for data mining, it is more difficult to do it. I am sure the prospect of data mining, or I will not build a forum that discusses commonly discussed. For the discussion of data mining prospects I once wanted to be a substantive thing of data analysis, now just a prototype, you can also find a reference in the forum. The following is a once in Sayuri and friends to discuss content: [Benpian full text] [Reply to this article] [Benpian Author: nohau] [Benpian Popularity: 160] Sender: nohau (on stilts birds), the letter area: Datamining Title: [Collection] Please give a recommendation to do network communication or data mining, more promising? Sending station: Nanjing University Small Lily Station (THU APR 15 16:23:48 2004) WeastWind (Western Wind) in WED APR 7 19:59:23 2004) mentioned: Please give a proposal to do network communication or data mining, more have a future? Network communication is now developing normal fire, what extent will the future network develop? What direction can do? Data mining is undoubtedly a very new research direction, can you develop in just a few years, form a fire like today? Do you give a suggestion? MiningBoy (Chai S-Tian, Nine-nine endless || Looking for a mine boy) in Thu Apr 8 09:20:37 2004) mention: Look at your own interest RSL (barbarian) in Thu Apr 8 09:30:33 2004) It is mentioned: I have done a data mining, enter the communication equipment manufacturing enterprise, useless nohau (Stepping on the stilt bird) in THU APR 8 09:42:47 2004) mentioned: If you refer to Huawei, If the company in ZTE, Lucent, Nortel, data mining is not completely useless, and now the stability of communication equipment is not very high. It is often a variety of problems in manufacturing companies like Nortel, Lucent. Huawei and ZTE are more exclusive, so the quality prediction of communication equipment is especially important. According to I know that there is a special group in Nortel to do the quality prediction of communication software, and publish a lot of articles, the effect is also good. If you want to continue excavation, quality forecast is also a good direction. RSL (Barbarian) in THU APR 8 11:52:06 2004) mentioned: ZTE Huawei did not see the use of DM, do 3G equipment, internal departments, did not see DM applications. NOHAU (Birds Stepping on the Stilt) in THU APR 8 13:03:19 2004) mentioned: So I only mention Nortel, Huawei and ZTE have not yet arrived at this level, Dequator (not spicy ~ I love south open) in Fri APR 9 12:46:11 2004) He mentioned: Isn't it that data mining current major application is financial, retail and communications industry? JUEWW (Sens) is mentioned in SAT APR 10 14:32:50 2004) mentioned: Good naive. . .
OneROAD (Mouth) was mentioned in Sun Apr 11 20:46:35 2004) mentioned: I feel that data mining and specific object relationship is very large, it is a very biased discipline in foreign countries, but it is temporarily in China. It's not very good after all, China's information is not very good. The data is not easy to get the credibility of additional data is also worth discusing, many of China, such as finance, communications, etc. their decision-making process is often A lot of government intervention, even if it is well received, it is not possible to do it. According to I know, the telecom department is developing a large data mining system, the investment is not small, but it is still a pilot, if it is not very successful, telecommunications The department is likely to abandon data mining this piece. So I feel that I have a network newsletter to be a little easehawking (with eagle hunting) in Sun Apr 11 23:02 2004) mentioned: Have A Look at Knowledge Plane, Which Uses Machine Learning Appoach To Control, Diagnose, Protect and Configure Internet Based ON Data from Nodes of the Network.juewW (Sens) in Mon Apr 12 13:02:49 2004) Refer to: This kind of thing, even if it is successful, it is useless. . . China's industries have only monopoly and no competition. DM cannot get really valued unless competitive is hot. NOHAU (Birds on the stilts) in Mon Apr 12 19:18:37 2004) mentioned: Xiongtai, this is a very confident, huh: JuewW (Sens) in Mon Apr 12 21:46:50 2004 ) Mentioned: No way. . . Who says that this kind of DM is so-called application prospects, I am uncomfortable. . . OneROAD (Mouth) in Mon Apr 12 21:52:52:18 2004) mentioned: The fact is that, the laboratory has received a few oil data mining projects to make a lot of things, until half a year knows everything It is nothing oil data at all, there is no petroleum data. The people of the Petroleum Bureau have said that the data is there as high in the two buildings. If you are guided to your computer, you will be almost the time you graduate. Point algorithm study Pap, casually research, it can be really sad, in order to invite a good job to the superior, he also hits high-tech people NOHAU (Stepping on the stilts) in Tue Apr 13 12:56:17 2004) mentioned: Data mining There are still many problems in practical applications. The research of data mining in these two years seems to be very hot, but the real mature application is small, most of the research in the algorithm is improved, and the data set in UCI is constantly Test their performance, and many research in order to display the data set for You Learner to compare the data set of Learner in order to display the superiority of its Learner, and remember the class of the last Longxing data mining. The clustering algorithm At the time, I also stressed the process of proposing a new algorithm: first, I want to see what kind of data set is defective in what kind of data set, because there is no algorithm to have the same good performance for all data sets, so You must find such a data set D, so that the clustering effect of others is very poor; then design a clustering algorithm according to D's characteristics can avoid the defects of others' algorithms on D, so a Paper is formed .
However, such paper has the development of data mining this discipline, there is no one-touch, data mining should be a very close discipline with the actual application, but the current research is as if it is trying to turn it into one. Aerial pavilion. In practical applications, the improvements of these proposed algorithms are not used at all, and they can display better performance only on the designated data set. Of course, no matter how foreign colleges are papers, in order to complete the task on time, you have to do some of these Paper, especially in our country, many school doctoral students cannot graduate from time to graduation, regardless of the pressing pressure. But as a Miner, I think it is necessary to consider this algorithm in the case of this algorithm. Is there anything (outside of Paper). Although the application of data mining has a long way to go, it is still unprinted, but our Miner should first lose confidence in it. If we are doing people think that they are doing, unruly, Real things, how do you think about this discipline? How do you think about us? Studydm (Hammer_shi) was mentioned in Tue Apr 13 13:49:05 2004) mentioned: I thought you really lost confidence. :( Individual is very optimistic about DM prospects. In fact, I can also find an authoritative support sound: A recent Gartner report lists five key technologies that will have important impact on industries in the next 3 to 5 years, of which KDD and artificial Intelligent ranking first. At the same time, this report will be parallel computer architecture research and KDD into 10 new technical fields that should invest in the next five years: broadband, wireless, Linux, content management, real-time analysis, data mining, security, Middleware, certification skills, business intelligence, and knowledge management. (I just know that ERP is also defined by Gartner) Why can't China can realize the company or industry DM and China? I think it is first as researchers. We should reflect on the problem. Perhaps our pioneers have done some hanging sheep's head to sell dog meat, let us fall into the embarrassment. Why can't we do our own efforts now? Change the status quo? A friend tells the case of oil, I think you are first accused! (Sorry, perhaps heavy, I am just right, please forgive!) According to my opinion, you don't have a data mining development. I always think that data mining is just a submodule of the decision support system (now only because people are eager to make them naked, now the DOS system during the OS development period), and the decision support system development requires system design and analysts. Industry expert leaders, the initial US users don't know what they need, and I don't know what conclusions he can draw. It is necessary to design and developer to make a raw model further and user discussion. This process repeatedly until users are satisfied. As for I don't think there is any problem. As a user, it can measure the workload. Maybe you are fearful to waste time for the paper. Developers don't confident, what do you have to monopolize in the Chinese industry? Status, I know that Anhui Mobile began to do data mining last July last year. Is Mobile and Unicom not competing nature? As for other industries such as insurance, retail, etc., mainly related to business privacy issues, but if the industry is competitive or data Safety is guaranteed (operated by large companies or internal employees) they are still useful. JuewW (feel) in Tue Apr 13 14:34:46 2004) mentioned: Poor. . . Another cheated good child. . . So the boss tells you the story of the foreground, the eyes enlarge, your boss's three-inch tongue, you can touch the so-called experts and connect the project. .
. Temptation temptation children are too easy. ^ _ ^ imcsee (imcsee) in Tue Apr 13 14:39:42 2004) mentioned: About the artificial intelligence affects the industry within 3 to 5 years? I can't see it, but I think that in a long period of time, AI needs to solve the industry. It is always solved:) JUEWW (Sens) in Tue Apr 13 14:41:44 2004) mentioned: You said that it is a truly thing to do things, not PAPER. . . In fact, the entire field of machine learning is like this. In many cases, data can only be picked up by one. . . However, the so-called data mining environment is not necessary. It is not necessary to buy someone else 's expensive software. The key is to do it. In addition, I see that the report of Gartner is hype, or the industrial taste is too strong, I didn't see anything different from DM and BI. . . OneROAD (Mouth) is mentioned in WED APR 14 03:30:30:58 2004) mentioned: In fact, our students still want to do something, but some things are really not we can decide, huh, wait, whether the company is only a teacher If you want to do something on your face, they don't want to give data at all. No matter how much time it gives them, the company does not want to spend money to organize the data. They are only given to the superiors, how much research funds are always in text. Used, otherwise, if you don't know how much, how much does it cost? This only costs people understand that anyway, they don't understand, they don't understand the country. There are still many things to be improved, not DM's problem, nor is the problem of researchers, not a problem, but a lot of aspects. This brother's opinion, I can understand, I also have the same feeling, but in a word of our teacher, "You are too young", huh, huh, some things really don't imagine, then simple DM is still in China At the beginning, the environment is still immature, I am afraid that banks, communication is a bit decent, which is CRM. 3 - 5 years later, it may be better, how to fry abroad, that is foreign, but the domestic is domestic, people's concept needs to be changed, this is a process. My brother has graduated, and she should hire a lot of units, including many big companies, such as Yahoo and more, but when the interview is reported, everyone has the feeling of listening to heaven, although many companies have to grab him, but that is not because DM. There are also some big companies in China (research institutes are not counted) to pay attention to DM, such as the square, but there is no climate, there are many small companies in Zhongguancun to do DM, but according to the brother, there is no research strength, too little cattle, Means nothing. DM is a very biased discipline, but the domestic information-based decision is still not enough. Foreign companies may still have a good job, and state-owned enterprises are really too many surface articles. Everyone is awake. Good. For the article, DM is still better, one is new, in addition, both PR, ML, AI, DB, can be cast, and it is more likely to pan, I have a core, simply, and Irrigation, Because there is not much person in China, the truth, the domestic DM is a bit specifically fried, especially in colleges, many teachers have DM's direction, but they really understand, there is not much, at least I understand Some teachers are really chasing the wind.
