Coggle 30 days of ML July competition learning

Part1 Content introduction

In the process of sharing knowledge , I found that many students have many problems in learning competitions ：

• Pandas、Numpy Not familiar with data processing

• Sklearn、LightGBM Not familiar with the model

• Don't know how to build Feature Engineering 、 Screening features

• NLP The model doesn't know how to build

The above questions are all about a contestant 、 An algorithm engineer must have . So we will organize a competition training camp this month , Hope to help you get started with the data contest . In the event, we will arrange specific competition tasks , Then, the students who participated in the class continued to break through the barrier to complete , It is possible to help you get started .

7 This month's competition will start with the following two competitions ：

• Diabetes genetic risk testing challenge

• Multilingual transfer learning challenge in the automotive field

Part2 Arrangement of activities

• Activities are free learning activities , There will be no charge .

• Please add the following wechat , And the reply 【 Competitive learning 】, To participate .

---

Part3 Points description and reward

In order to motivate the students to complete their learning tasks , Divide the learning tasks according to their difficulty , And according to whether the task with high and low scoring difficulty is completed or not, the scores are 3、2 and 1. At the completion of 7 After monthly study （ This activity , By the end of 7 month 24 Number ）, The evaluation will be conducted in the order of points Top3 learners .

Punch cards can be written in an address , Each time there is a new completion, you can repeatedly submit and punch in ！

Top1 Of learners will receive the following Reward ：

• Coggle Competition interview opportunity

• 《 Machine learning algorithm competition practice 》

Top10 Of learners will receive the following Reward ：

• “ IFLYTEK x Datawhale” The certificate of outstanding player jointly issued .

• Coggle Fringe benefits

• Coggle Competition interview opportunity

Part4 Diabetes genetic risk testing challenge

Learning content

This course is mainly aimed at the diabetes genetic risk detection challenge , It will explain the specific knowledge points and details used in the data competition . In this study, we will learn feature engineering 、 Feature selection and model parameter adjustment process .

The address of the competition registration ：http://challenge.xfyun.cn/topic/info?type=diabetes&ch=ds22-dw-gzh01

Introduction to the contest question

In this competition , You need to build a genetic risk prediction model for diabetes through training data sets , Then predict whether the individuals in the test data set have diabetes , Join us to help diabetes patients solve this problem “ Sweet troubles ”. For individuals in the test data set , You must predict whether they have diabetes （ Have diabetes ：1, No diabetes ：0）, The predicted value can only be an integer 1 perhaps 0.

Training set （ Game training set .csv） Altogether 5070 Data , Used to build your forecasting model （ You may need to do data analysis first ）. The fields of the data are numbered 、 Gender 、 Year of birth 、 Body mass index 、 Family history of diabetes 、 diastolic pressure 、 Oral glucose tolerance test 、 Insulin release test 、 Triceps brachii skinfold thickness 、 Signs of diabetes （ The last column ）, You can also use feature engineering techniques to build new features .

Test set （ Competition test set .csv） Altogether 1000 Data , Used to verify the performance of the prediction model . The fields of the data are numbered 、 Gender 、 Year of birth 、 Body mass index 、 Family history of diabetes 、 diastolic pressure 、 Oral glucose tolerance test 、 Insulin release test 、 Triceps brachii skinfold thickness .

Punch in summary

The name of the task	difficulty
Mission 1： Sign up for the competition	low 、1
Mission 2： Game data analysis	low 、1
Mission 3： Logistic regression attempts	low 、1
Mission 4： Feature Engineering	in 、2
Mission 5： Feature screening	in 、2
Mission 6： High order tree model	in 、2
Mission 7： Multi fold training and integration	high 、3

Punch in requirements

notes ：

• Need all the tasks to be written in one blog

• It is recommended to add a thought process to the punch in process , You can join in trying & Data record

Part5 Multilingual transfer learning challenge in the automotive field

Learning content

This course is mainly aimed at the multilingual transfer learning challenge in the automotive field , It will explain the specific knowledge points and details used in the data competition . In this study, we will learn text classification and keyword extraction .

The address of the competition registration ：http://challenge.xfyun.cn/topic/info?type=car-multilingual&ch=ds22-dw-gzh05

Introduction to the contest question

In order to improve the competitiveness of products, domestic automobile enterprises 、 Better go to overseas markets , Put forward the demand for intelligent interaction in overseas markets . But countries around the world are “ Data security ” There are strict legal restrictions on , Do a good job in overseas intelligent interaction , The biggest challenge for local enterprises is the lack of data . This competition requires the contestants to pass NLP Relevant artificial intelligence algorithms to achieve multilingual transfer learning in the automotive field .

In this transfer learning task , IFLYTEK smart car BU There will be more in car human-computer interaction Chinese corpus , And a small amount of Chinese and English 、 China and Japan 、 Chinese and Arabic parallel corpora are used as training sets , Contestants build models from the data provided , Carry out intention classification and key information extraction tasks , The final use of English 、 Japanese 、 Test and judge in Arabic .

1. Preliminaries

• Training set ： Chinese corpus 30000 strip , Chinese and English parallel corpora 1000 strip , Chinese and Japanese parallel corpora 1000 strip

• Test set A： English Corpus 500 strip , Japanese Corpus 500 strip

• Test set B： English Corpus 500 strip , Japanese Corpus 500 strip

2. The rematch

• Training set ： Chinese corpus is the same as the preliminary contest , Chinese Arabic parallel corpora 1000 strip

• Test set A： Arabic corpus 500 strip

• Test set B： Arabic corpus 500 strip

1. This model is based on the submitted result document , use accuracy Evaluate .

• Classification of intention accuracy = The correct number of intentions / Total data

• Key information extraction accuracy = Number of critical information completely correct / Total data

Punch in summary

The name of the task	difficulty
Mission 1： Sign up for the competition	low 、1
Mission 2： File reading and text segmentation	low 、1
Mission 3：TFIDF And text classification	low 、1
Mission 4： Regular expressions	in 、2
Mission 5：BERT Introduction to models	in 、2
Mission 6：BERT Text classification	in 、2
Mission 7：BER Entity extraction	in 、2

Punch in requirements

notes ：

• Need all the tasks to be written in one blog

• It is recommended to add a thought process to the punch in process , You can join in trying & Data record

One key, three links , Learning together ️