Tag:machine learning

  • Derivation of Logistic Regression from Principle-Logit Transform and Potential Factor Error

    Time:2019-8-24

    The application of Logistic Regression (LR) and the idea of engineering are generally introduced very clearly. Most methods start with Sigmoid function. This blog post attempts to reinterpret how LR is derived from other perspectives. Logit transformation For predicting a categorized variable, a common way to generalize OLS is to use it directly. $$P(y = […]

  • Machine Learning (6) – Realization of Automatic Subject Type Partition of Movies Based on KNN Classification Algorithms

    Time:2019-8-23

    Introduction to Classification Algorithms As we all know, movies can be categorized according to the theme, but how to define the theme itself? Who decides which theme a movie belongs to? That is to say, what public characteristics do movies with the same theme have? These are the issues that must be considered when classifying […]

  • Preview of the New Book “The Gate of Full Stack Data”

    Time:2019-8-23

    Finally, I can give you all the relatives and friends who care about “The Gate of Data on the Whole Stack” a confession! After more than three months of editing and typesetting, the final edition has been completed. After New Year’s Day, when you come back after eating, drinking and playing, you can start printing. […]

  • SLS Machine Learning Best Practice: Batch Sequence Anomaly Detection

    Time:2019-8-22

    1. High Frequency Detection Scene Scenario 1.1 Cluster has N machines, each machine has M time series indicators (CPU, memory, IO, traffic, etc.). If we model each time series curve separately, we need to write too many repetitive SQL by hand, and the computing cost of the platform is very high. How to better use […]

  • Transfer Learning in NLP

    Time:2019-8-21

    Summary:Transfer learning is developing in all fields, and the NLP field is being impacted! In our previous articles, we showed how to use CNN and migration learning to build classifiers for our own images. Today, we introduce the latest trends in migration learning in NLP and try to categorize the data sets reviewed by Amazon […]

  • Implementation of Machine Learning Decision Tree in Sklearn

    Time:2019-8-20

    Implementation of Decision Tree in Sklearn Hello, ladies and gentlemen.( ̄▽ ̄) Well, first of all, let me state that my development environment isJupyter labThe libraries and versions used are for your reference. Python3.7.1 (Your version should be at least 3.4 or more) Scikit-learn0.20.0 (your version should be at least 0.20) Graphviz0.8.4 (No decision tree can be […]

  • Detailed Explanation of Important Parameters of Classified Decision Tree in SKlearn

    Time:2019-8-19

    classsklearn.tree.DecisionTreeClassifier(criterion=’gini’, splitter=’best’, max_depth=None, min_samples_split=2, min_samples_leaf=1, min_weight_fraction_leaf=0.0, max_features=None, random_state=None, max_leaf_nodes=None, min_impurity_decrease=0.0, min_impurity_split=None, class_weight=None, presort=False) 1. Important parameters: criterion In order to transform a table into a tree, a decision tree needs to find the best node and the best branching method. For a classification tree, the “best” index is called “impurity”. Generally speaking, the lower the […]

  • When in-depth learning meets automatic text summary

    Time:2019-8-19

    Welcome to Tencent Cloud + Community to get more Tencent Massive Technology Practice Dry Goods~ This article was published by columneditor in the Cloud + Community column Author: Yao Junlin Introduction: With the explosive growth of text information in recent years, people can access a large amount of text information every day, such as news, […]

  • The basic idea of parameter adjustment in machine learning

    Time:2019-8-18

    I found that most of the books related to machine learning are traversing all kinds of algorithms and cases, explaining the principles and uses of various algorithms for you, but little research has been done on parameter adjustment. There are many reasons for this. One is that the way of parameter adjustment is always based […]

  • A More Effective Method of Transfer Learning in NLP

    Time:2019-8-17

    Summary:Transfer learning is widely used in the field of computer vision, while the field of NLP has just begun. This paper introduces two kinds of transfer learning methods in NLP field. They are the embedding layer of training and the fine-tuning model. At present, the embedded layer is widely used, but the fine-tuning model comes […]

  • Implementation of Random Forest in Sklearn

    Time:2019-8-17

    Hello, ladies and gentlemen.( ̄▽ ̄) I am a vegetable, my development environment isJupyter labThe libraries and versions used are for your reference. Python3.7.1 (Your version should be at least 3.4 or more) Scikit-learn0.20.0 (your version should be at least 0.19) Numpy 1.15.3, Pandas 0.23.4, Matplotlib 3.0.1, SciPy 1.1.0 1 Overview Overview of 1.1 Integration Algorithms Ensemble […]

  • TO-DO-LIST in Data Mining

    Time:2019-8-17

    Data Mining Process and Method 1. Tasks: association analysis cluster analysis Classification analysis Anomaly analysis Analysis of Specific Groups Evolution Analysis 2. Methods: Statistics On-line processing analysis information retrieval machine learning classification Practical Application: Application Classification/Trend Forecasting/Recommendation of Related Goods regression analysis Practical Application: Predicting Sales Trends clustering Practical application: classification Association Rules There are […]