[live online] text classification technology in artificial intelligence

Lecturer: Huang Hongbo

Cdn.xitu.io/2018/9/28/1661ee5b4d8c4144?imageView2/0/w/1280/h/960/format/webp/ignore-error/1 ">

Instructor profile:

Zhuhai Jinshan Office Software Co., Ltd. (WPS) artificial intelligence expert, senior algorithm engineer, has many years of software development experience, author of "TensorFlow Advanced Guide fundamentals, algorithms and applications". He has worked as an expert in the field of artificial intelligence in big data Center of Gree Electric Appliances Co., Ltd., and has served as senior engineer, technical manager, technical director and other positions in many companies. He has led the team to develop intelligent payment system, recommendation system and intelligent question answering system based on face recognition technology. Good at data mining, machine learning, mobile development and other professional areas, and has a wealth of practical experience.

Sharing outline:

1. Common misunderstandings in text classification practice:

1.1 depth model must be better than traditional machine learning model.

1.2 if the accuracy is high on the verification set, it is high on the line.

1.3 accuracy is the most important index in text classification

1.4 the more complex the model, the better the effect.

two。 Text classification strategy

2.1 Select the appropriate training set and test set

2.2 skillful use of word bag model

3. Comparison of commonly used text classification models

3.1 comparison of machine learning algorithms

4.1.1 LR+ word frequency

4.1.2 LR+one-hot

4.1.3 Standardization + word Frequency + LR

4.1.4 regularization + word frequency + LR

4.1.5 Laplace smoothing + word probability + LR

4.1.6 Bayesian + word frequency

4.1.7 Bayesian + feature extraction

3.2 comparison of deep learning models

3.2.1 TextCNN

3.2.2 textRNN

3.2.3 textRCNN

3.2.4 HAN

Audience benefits:

1. Understand some common pits in the process of text classification

2. Understand the trade-off of various text classification models in practical projects.

3. Understand the real data comparison of various algorithms under the same data set.

Event time: 20:00-21:30, October 16, 2018

Event details: [live online] text classification technology in artificial intelligence

Note: log in and register successfully immediately. After receiving the registration message, enter to add the live broadcast assistant Wechat, indicating that the "10.16 live broadcast" assistant will pull you into the group and inform you of the live broadcast channel before the live broadcast begins!

Follow the official account of [Pegasus], reply to the corresponding keywords to package and download learning materials; reply to "join the group", join the AI, big data and project manager learning group of Pegasus, and grow up with excellent people!