Motivation

traditional different evaluation parameters are an average over parts of the distribution of impacts:

• The ATE averages over the entire distribution
• The ATT averages over the distribution of impacts for those who are somehow allocated into treatment
• LATE averages over the distribution of impacts for those who switch into treatment as a result of a reform or more precisely, as a result of a change of the value of some instrument a§ecting decisions to participate.

What is the difference between a function decorated with @staticmethod and one decorated with @classmethod?

the answer is copyed from stack overflow

Neural Network Tutorial

This note introduce the basic concept of neural network in machine learning, plus the python realization. this note is intended to tell beginners the cornerstone of the neural network and help anyone who is interested in neural network can build their own application easily.

The notes contains the following part:

• Toy example of Neural Network
• The Defination of Neural Network
• Bsic element in Neural Network and Backprogation
• constuct the Neural Network

With the help of Navicat for Mysql, we can easily transform different data format into mysql. And when we install MySQL Workbench, it helps us to deal with any sql operation easily and conviently.

get column name from table

1

select column_name from information_schema.COLUMNS where table_name='table name'

python 上面有不少数据可视化的工具包，先利用这篇文章总结如下：

安装 pip

1
2
3
4
5
6

later use utf-8-sig instead of utf-8

pip install --target=d:\somewhere\other\than\the\default package_name
带空格处理
pip install --target=X:\My~Documents\py_pkg\ selenium

Here I summarize some matlab trips to facilate my research

This is a few summary of my economics studying notes in my graduate and undergraduate

In this article, I sumamrize some selenium webdriver api command to help coding webspider.

ECONOMICS

Information Asymmetry and Bidding Behavior in Common Value English Auction

Abstract: In common value English auctions, bidders may have different level of private information toward the selling item. Studying the such information asymmetry requires proper estimation of the value distribution.However, the same distribution of bids can be rationalized by many value distributions. Moreover, the same value distribution may lead to many different bid distributions due to multiple equilibria and model incompleteness.This paper overcomes the “many-to-many” correspondence utilizing moment inequalities implied by the bidder’s bidding history and develops a structural econometric model to estimate the value distributions. I use this structural model to study the effect of information asymmetry on agent’s bidding behavior. In addition, this paper proposes a parametric approach for testing the information structure. The paper finds that the information premium mainly comes from the informed bidder’s screening effect and is independent of the number of participants. Applying the data from Chinese Justice Auction, I find that the noisy part in private signal is very large. The large noisy part reduces the selling revenue.