Notes on the startup of the oldest programmers: full-text search, data mining, and recommendation engine application 28

Investment Portfolio
Although there is almost no progress in project financing, Wu Yan found that the number of registered users of the website is still increasing, and now it has already exceeded one thousand, and there will be more than a dozen registered users every day, in addition, dozens of users have become active users, regardless of the number of logins, number of browsing products, and number of writing comments, they are steadily increasing. Some people are even writing product blogs, and some users are already revising the content in the product WiKi. These should be not search engine spider, but actually users, wu Yan is very pleased to see that his website has gradually become a real user.
Wu Yan thinks it's time to consider adding a collaborative filtering mechanism, but the first step is to analyze existing users, because every time they communicate with VC, they will ask who your target customers are, what are their characteristics? Every time Wu Yan's answers to this question are somewhat unsatisfactory. Now there are hundreds of user data, which seems to be enough to be mined.
All the required data entered by the user during registration includes gender, age, occupation, income, and other information. The data is used as the user's attribute and the online time is counted, we should be able to find some regularity. Wu Yan started to implement this idea. Wu Yan first found out the users with complete information and made a preliminary statistics. About 300 users were found, and Wu Yan randomly took 200 samples as training samples, the remaining 100 samples are used as test samples. Then, the WEKA framework is used to define user attributes, input training sample data, and the RBF Network is used as the prediction model to run the prediction model, then, a test sample is run to evaluate the prediction model. The prediction model is basically available based on evaluation criteria such as mean variance and average absolute error. In this way, Wu Yan can predict the online time of a new user based on gender, age, occupation, income, and other information, this makes sense for future promotions. Through these algorithm analysis, Wu Yan also found some characteristics of his active user group, such as male, between 25-35 age, and white-collar occupation, this is not consistent with the active user groups of other e-commerce websites. Wu Yan guessed that there may be two reasons: first, the website is dominated by 3C products, men are more interested in these products, and their websites have not been widely promoted. Currently, most of the users are in the industry, while most of them are in the industry.
After completing these functions, Wu Yan is uploading code and documents to his open-source project on the Sourceforge.net website. Unexpectedly, although this project is still not complete, someone has started to download the trial. Will it be another competitor? Wu Yan is always curious about this.
On this day, Wu Yan received an email from Zhao Haibo of Qingwei venture capital. The subject of the email is "Fw: Re: health portal business intelligence application". Sorry, this topic is very strange, wu Yan could not think of any relationship with himself, but Zhao Haibo, vice president of Qingwei venture capital, could have sent an email to himself. Perhaps he had an eye on his financing. He was excited when he thought of Wu Yan here, you can't wait to open the email.
Email is an internal discussion email about the health portal. The basic background is that although the health portal "Health 120" currently has a lot of medical resources and a lot of professional medical content, at present, I have made many attempts, such as adding SNS, Q & A modules, and even trying pharmaceutical e-commerce, but the results are not satisfactory. After analysis, I think the crux is that there is no good personalized function, at present, all the attempts are focused on adding functions to the perimeter of the portal. Because the portal style cannot be separated, SNS and Q & A functions can only be configured. At present, we need to introduce Web2.0 elements to make full use of medical resources and develop new products and services.
Zhao Haibo wrote this in his email to Wu Yan: Wu Yan was very impressed with your last communication. "Health 120" is a health portal project we invested in, below is an internal email memo from them. They are currently preparing to develop brand new Health Community products, but they do not have the corresponding technical accumulation. We want to hire you as our industry expert, I have discussed this technical issue with the Health 120 team. Do you have time to participate?
Wu was originally invited to participate in technical exchanges. Wu was a little disappointed, but he thought that Zhao Haibo valued himself so much that he could be an industry expert and participate in their project discussions, wu did not hesitate to agree.
On the day of the meeting, Wu Yan almost arrived at Qingwei Capital One hour ahead of schedule. As soon as he arrived at the front-end, the front-end mm looked up and said, "Hello Mr. Wu !"
Although Wu Yan has been here for a while, he recognized himself at the front-end and said hello to himself, which surprised Wu Yan. At the same time, it is strange that Wu Yan believes that he is not important enough to let the front-end mm remember himself at once. How can the front-end mm have such a good memory?
The front-end mm politely said, "You have been here for an hour! Mr. Zhao has also arrived. Please sit down in Zhao's office for a while !"
Wu Yan hurriedly said, "No, it's not a good customer. I will wait here for a while and I will not join Zhao Zong !"
The front-end mm then said, "Zhao told me just now that if you come early, you will go to his office for a while ."
After hearing this, Wu entered Zhao Haibo's spacious office. Zhao Haibo saw Wu Yan come in, pushed his notebook next to him, and put Wu Yan on the sofa. politely said, "Sorry, please come to our technical seminar this time !"
Wu Yan hurriedly said: "Zhao Zong, you are welcome! I think it is good to attend your technical seminar !"
Zhao Haibo and Wu Yan both sat on the sofa and handed a piece of information to Wu Yan, saying, "This is the next phase of the product R & D plan for the health 120 portal. You should familiarize yourself with it first, the contents will be discussed later."
When Wu Yan looked at the materials, Zhao Haibo went on to say, "I will introduce you to the basic information first. The health 120 Portal is a vertical portal for our investment in the past few years, at that time, I thought that with the Health 120 team's medical background, health care is a hard demand with the improvement of living standards, and that the prospects are bright."
Zhao Haicang paused and said, "But over the past few years, we have found that our business has never improved, because we signed a gambling agreement with the project team, we have gained a vast majority of health 120 equity, which is of course not our hope. The biggest fear of VC is the final shareholder of investment. In this way, entrepreneurial teams are becoming less and less motivated, at the same time, it is increasingly difficult for us to exit, so we need to find a solution to this problem. Of course there are other problems in the project, but this is the biggest problem currently ."
Wu nodded, and Zhao Haibo said, "so now I want to transform into a medical community and Q & A products and services ?"
Zhao Haibo nodded and said, "This is a plan developed by their project team, hoping to get our next round of investment. However, we are cautious at the moment, so please come and discuss it together to see if this direction is feasible. I mainly want to know whether the full-text search, data mining, and recommendation engine technologies in your project can be applied to the health field ."
Although this was Wu Yan's first attempt in the health field and the first time he thought about the application of full-text search, data mining, and recommendation engine in the health field, Wu Yan thought a little bit and had a lot of ideas, "I think Data Mining is widely used in the health field. For example, users can enter gender, age, medical history, and other information on the website, if we can also record some habits, we can develop data mining algorithms to determine which groups are vulnerable to disease and track the side effects of medication, users who can determine the gender, age, and medical history are most likely to receive side effects using a specific drug."
Wu Yan has many other ideas, but for a moment, they are limited to the limitations of expression ability and cannot be listed one by one. After thinking about the ideas, he went on to say, "for example, when users choose to take medication, we can make recommendations based on our age, gender, and medical history, and consider the cost and safety factors, you can also see other users who are similar to the situation after using the drug ......"
Wu Yan spoke passionately and felt that if he had given himself more time, he could say a lot.
Zhao Haibo nodded when listening to Wu Yan and said, "You are very right. As long as we have solved some policy risks and ensured medical security, these fields are indeed promising !"
Yes. The most important thing for a healthy website is medical security. Why didn't you think of it! It seems that Mr. Zhao's experience is still the same, and he suddenly found the key to the project. When thinking of this, Wu Yan couldn't help but admire Zhao Haibo.
"Time passes so fast. We will soon start. Let's have a meeting first !" Zhao Haibo looked at the table and said.
Zhao Haibo and Wu Yanyi entered the big meeting room of Qingwei venture capital. On the other side of the meeting table, it is estimated that they were members of the Health 120 team. They arrived early. Qingwei venture capital was here, two Investment Managers of Qingwei venture capital have also been employed.
When Zhao Haibo came in, he stood up and greeted everyone. Zhao Haibo greeted everyone and introduced Wu Yan: "This is Wu Yan, an industry expert in the field of business intelligence !"
Wu said hello to everyone, and the meeting began.
First, the CEO of the Health 120 team introduced the next phase of the plan. The PPT was written by Zhao Haibo gang to Wu Yan. The CEO's speech skills were good, not only did he clearly introduce the project, it is also very inflammatory, and Wu Yan is not as good as himself.
After talking about the health 120 CEO, Zhao Haibo summative said, "That's good. Thank you, Mr. Wang for sharing the product plan for the next stage. What do you think ?"
At this moment, an investment manager sitting next to Zhao Haibo first spoke and said, "There is no problem with the project's general direction, but I personally think there are some problems with the project's budget ." He glanced at Zhao Haibo and seemed to have a positive answer. Then, he said, "I think it seems too big to invest 1.5 million yuan in the development of this new business. In addition, 3 million of the advertising cost during the project promotion stage is also high ."
"Yes," replied health 120ceo. "We need to build a new R & D team. The cost of data mining experts is very high. We plan to introduce three of them, each person costs 30 thousand RMB per month. In addition, we need three cloud computing experts who need to be proficient in hadoop technology, and each person needs 30 thousand RMB in cost. In addition, we need six developers, such a team needs to spend at least half a year on development, with a cost of at least 1.5 million. As for the advertising cost, we had to pay more than 120 for the early health 3 million portal. Considering inflation, this part of the cost should not be lower ."
The two sides seem to be about to argue, and the atmosphere is a little tense.
Zhao Haibo said at this moment: "Good! Let's take a look at the R & D cost. Let's first listen to the opinions of industry experts. Let's talk about it !"
Wu Yanqing cleared the noise and said: "for R & D, I don't think we can use three experts. Only one expert is needed, at the same time, the cost of 30 thousand million RMB per month is also higher. In terms of cloud computing hadoop, I think a technical master is almost the same. He can take charge of controlling the MAP/reduce algorithm design, it is also responsible for hadoop configuration and performance optimization. In addition, it is best to spend about three months in Project Phase I so that user feedback can be collected as soon as possible to avoid useless work ."
Zhao Haibo waited for Wu yigang to finish, and then said, "Yes, Wu Yan makes a lot of sense. You should go back and think about it carefully !"
After a pause, Zhao Haibo went on to say, "When it comes to advertising fees, the original advertisement contract you signed should not be over. I suggest you add new product elements to subsequent advertisements, in this way, advertising fees can be greatly reduced."
Health 120 teams face each other and don't know how to respond at a time.
Zhao Haibo went on to say: "There is a new website in the health community, called the health care community. What do you think of this website ?"
"We have not studied this website carefully !" The CEO of Health 120 portal replied.
Zhao Haibo said with great enthusiasm: "You should take a good look at them. Their business has been very good. We have carefully studied them. I believe you will know better than us !"
The team members of the Health 120 portal did not know how to answer the question at the moment, and they were a little embarrassed.
Zhao Haibo glanced at them and said, "but you don't have to worry about this, because we found that their user information is theft of the user information of our health 120 portal! So we still have a solution ."
"But the user information is similar and common! It is difficult to find evidence to steal user information !" Health 120 portal CEO reminded.
Zhao Haibo smiled and went on to say, "That's good. We usually have no way to do this, but this time we have made some preparations, because we have added hundreds of fake user data to over million user data, we found that these fake user data exists in their user data, they can hardly explain this! We have already asked our lawyers to prepare a letter of attorney and applied for evidence preservation from the court. This should be okay !"
At this moment, the Health 120 portal team was silenced. Wu Yan was confused and felt that the technology exchange was somewhat different. However, Wu Yan still couldn't guess what it was.
Zhao Haibo then said in a relatively relaxed tone: "Of course, we do not have to resort to the law. If we agree to let us hold our holdings, we can also work with the medical care community. What are your opinions ?"
The Health 120 portal team nodded and reluctantly said, "Yes, yes !"
The meeting ended like this. After the meeting, Zhao Haibo invited Wu Yan to his office. After the two men sat down, Zhao Haibo first opened his mouth and said: "do you think today's Technical Exchange conference is a bit strange."
"Yes, yes !" Wu Yan has always been very strange.
"Yes," Zhao Haibo said slowly. "due to the previous gambling agreement, there are currently less than 120 of the total shares of the founding team of health 12% portal, so they must have some ideas, these are understandable. However, on the one hand, they imagine that today, we have invested so much in our minds that we don't want to get involved, and we are constantly asking for money from New Directions. On the other hand, they are using the user data of the Health 120 portal, I secretly joined the medical care community. Fortunately, we added fake data to user data, so that we found that they wanted to create another portal, so we had just discovered the scene!"
Wu Yan suddenly realized that this was a banquet, which was originally presented to the Health 120 portal team and wanted to take back what they had done in private, wu Yan could not help but secretly admire Zhao's brilliant position.
Zhao Haibo went on to say, "This time I called you here. Besides discussing technical aspects with them, we also think that the Health 120 portal management team has a big problem, in particular, our CEOs and ctos are ready to replace them. We have already selected the candidates for the CEO. the CTO is difficult to choose because of our background, so think of you, check whether you are interested!"
This was the case. Wu Yan finally understood it at the moment. CTO, This Is What Wu Yan previously dreamed of. Now he suddenly appeared in front of himself. How can he not let Wu Yan feel lost! Moreover, Wu Yan's project has now reached the point where it is difficult. If he accepts this position, all problems can be solved. What a great opportunity can he give him.
When Wu Yan was thinking about connecting to the job, Zhao Haibo went on to say, "Our idea is like this, you and our designated CEO first enter health 120 as ordinary employees, and then squeeze out the existing health 120 portal management team after a series of operations."
Wu Yan couldn't help but fight a Cold War when he heard these things. Such things appeared only in TV series in the past and will be staged in reality immediately. Moreover, he must assume such a key role by himself, it's really exciting, but the key question is, are you doing well?
Zhao Haibo seemed to have understood Wu Yan's doubts and said, "You can rest assured that our CEO candidate is the current front-end. I am very confident in her, she will arrange everything later."
After stopping, Zhao Haibo went on to say, "Don't underestimate her and give you a detail. When you come in, does she recognize you all at once, do you know how she did it?"
Yes, Wu Yan had this question when he first entered the company. It was just because he had been in a meeting and forgot it. Now Zhao Haibo raised it again, and Wu Yan's curiosity was hooked up again.
Zhao Haitao proudly said: "In fact, she has developed a system and takes photos of every visitor. When the visitor needs to make an appointment for the second time, she knows the photos of these visitors in advance. In this way, when these people come back, of course she can recognize the visitor at a glance! We are very confident in the front-end who can think of this !"
In the past, Wu Yan's questions finally solved, and he also admired the front-end MM. But Wu Yan still thinks, how can she accomplish Zhao Haibo's arduous task?
Zhao Haibo seemed to have read Wu Yan's question and went on to say, "are you worried about how she could accomplish this? You don't have to worry about this. First, the Healthy 120 team doesn't know our front-end team, so we won't be wary of her, and then we will give them another feeling, that is, the small MM can communicate well with us, and she can easily be reused. At the same time, their team is not an ironclad team. It just forces us to take the CTO and then the CEO ."
This plan seems very feasible, but it seems a little strange. Although it cannot be said that it is not sinister, it also seems a little inactive. Wu Yan secretly thought about it.
"How? Are you interested in joining the health 120 portal management team in the future ?" Zhao Haibo asked.
Wu Yan was hard to make a decision when he was suddenly asked about his thoughts. However, Wu Yan knew that he had to make a decision while looking at Zhao Haibo's expectations. Wu Yan thought carefully and finally made up his mind and firmly said, "Thank you, Mr. Zhao for your kindness. I think I can do the same technology. I think I can't do this task as you said!"
I thought Zhao Haibo would be very disappointed after listening to it. Unexpectedly, Zhao Haibo nodded and said with deep understanding: "Yes, you are not suitable for doing this. The most valuable thing is that you know what you can do and what you can't do ."
Wu Yan thought that it was a waste of time to say goodbye to Zhao Haibo. But how can we make better use of this opportunity? Wu Yan is counting in his mind.
Coming soon: MBA students