Choosing a UNIQUE pattern classification approach and implementing that approach using one of the machine learning Methods in Python/MATLAB/Weka.
Structure investigation using the following steps:
Normalization, outlier detection, censoring of bad data, etc.
Handling of missing data, records of varying length, etc.
To generate many features from the time-series data provided and ultimately only use a subset inclassifier.
Partition data & establish experiment design
Train/validation/test sets, balancing classes (optional), etc.
What approach used, what parameters required, how they were tuned, etc.
Testing & expected accuracy
What is predicted accuracy, how was it computed, provide a standard error / standard deviation on your estimate (e.g. "the minimum of my 3 accuracies will be 0.73 ± 0.04")
Implement at least one meta-learning strategy (e.g. CME-voting, bagging, boosting), and investigate its effect on accuracy.
37 фрилансеров(-а) в среднем готовы выполнить эту работу за $521
Hi, I am Ibrahim, and I am a data scientist, I have expertise in machine learning, and python, I can create a machine learning algorithm, but first I need to see the dataset. Regards, Ibrahim Anjum
Hi I had a PhD in computer science. My area of expertise is data mining. I had worked on many projects using Weka. I think I can help you. Please contact me. Many thanks.