Project Experience

  • 1. Big Data Platform about Intelligent Diagnosis for Industrial Construction
    • Field : Machine Learning
    • Details :
      • Data preprocessing includes data normalization, missing value processing, data visualization and exploratory analysis.
      • Performed feature engineering on industry building vibration data.
      • Built event recognition model using machine learning algorithms such as random forest, gradient boosting tree and neural network.
  • 2. Industrial Accident Cause Analysis System
    • Field : Deep Learning, Computer Vision
    • Details :
      • Performed LSTM and attention mechanism model over the industrial data with time series information, find the time period which has the greatest relationship with the cause of the accident and the abnormal parameters, so as to improve the detection performance of the cause of the accident.
  • 3. Evaluation and Improvement of UAV Operator’s Maneuverability
    • Field : Machine Learning
    • Details :
      • Employed Python to crawl user flight log data and perform data preprocessing, including data cleaning, outlier processing and data visualization.
      • Performed Ant colony algorithm to obtain the optimal flight path for the given flight area.
      • Used the obtained features from the feature engineering, a GBDT model is performed to evaluate pilots’s flight ability.
  • 4. Supply Chain Revenue Management System
    • Field : Deep Learning, Optimization Methods
    • Details :
      • Implemented optimization method to reduce supply chain cost, select better offline warehouse address and reduce company operation cost.
      • Used the deep learning model to predict the offline customers’ purchasing situation to reduce storage costs.
  • 5. Stock Market Forecasting and Consumer Credit Modeling
    • Field : Deep Learning, Machine Learning
    • Details :
      • Analysed the stock market trends. Based on the business knowledge of financial experts and stock market experts, and combined with LSTM algorithm, a stock market rise and fall prediction model is established.
      • Combined market information, user personal information and user consumption situations to establish a credit model.