It is already 2020, and many students are thinking about the career development path. Recently, there have been many consultations with Mr. Chen. A very common question is: How long can the red flag of a data analyst be played? Focus on answering today. It is still the consistent style of Mr. Chen, not to brag and not to criticize, but to talk about things Norway Phone Number objectively. Let’s take a look at the five most tangled questions for data analysts in 2020: Question 1: Will data analysis be replaced by artificial intelligence? Answer: No! First of all, seeing all the titles that mention the word "artificial intelligence", everyone can think that it is a Norway Phone Number pseudo-question that causes anxiety. Those who really know how to do it all say: algorithms, machine learning, or simply plug-in CV, NLP, recommendation and other specific fields. Second, these two things are two things at all. The development of algorithms will not replace data analysis, but will make data analysis easier.
Why? Because, by nature, algorithms fight inefficiencies. Through the cycle of manual labeling-model training-prediction testing, algorithms can replace a lot of labor-intensive work in the past. However, there are preconditions for the algorithm to achieve this purpose: first, it needs a clear result: manually annotated graphics, credit default/non-default records, and so on. Second, a large amount of feature data is required for training the model. It is because Norway Phone Number of this that we see the most successful application of algorithms in the field of CV. The comparison of faces and documents is rich in features and the results are clear. There have also been corresponding Norway Phone Number progress in traditional risk control, recommendation and other fields. The progress of the NLP field is relatively slow, which is tossed by complex contexts. Essentially, data analytics fights uncertainty. When we want to analyze the problem, it is more: No data.
The new business has just been launched, and data collection has not been paid much attention in the past Fake data: business manipulation, lack of process, profit-driven Messy: inconsistent caliber, irregular process, and nonsense use Human distortion: business Norway Phone Number side is talking nonsense in order to keep KPIs Can't judge: only know how to write the year-on-year ratio, can't interpret the meaning Subjective assumptions: Doing data only to prove that you are Norway Phone Number right, arguing At this time, it is equivalent to no labeling at all, or a few artificial labels, and the model cannot be trained at all. Also touch a fart. That's why I have five big questions about data analysis: What is (quantify results, get data) How much is it (set standards, evaluate good or bad) why (find reasons, test assumptions) What will happen (comprehensive assessment, make trade-offs) So what (predicting prospects, exploring possibilities)
If you want to buy phone number list you can write here
1 post • Page 1 of 1