This book addresses the basic principles of statistical inference. The first chapter recalls the basis of probabilities. The second chapter is devoted to point estimation, the hypothesis test, including p-values, the confidence region determination, mentioning only the most important concepts. The main practical approaches, such as the least squares minimization, the moment method and the likelihood maximization, are broadly detailed. The following chapter is devoted to hidden Markov models (HMM) which are the basis of the more recent algorithms in signal and image processing. The final chapter is devoted to Monte-Carlo methods, which provide useful estimation tools using well-chosen random generators. To understand a theory well, it must be supported by practical examples. Therefore, 16 computational examples and 62 computational exercises using a real dataset are proposed. The solutions are coded in Python language and provides in an appendix.
1. Useful Maths.
2. Statistical Inferences.
3. Inferences on HMM.
4. Monte-Carlo Methods.
5. Hints and Solutions.
Maurice Charbit is Professor at Telecom ParisTech, France. He is a teacher in probability theory, signal processing, communication theory and statistics for data processing. With regard to research, his main areas of interest are: (i) the Bayesian approach for hidden Markov models, (ii) the 3D model-based approach for face tracking, and (iii) processing for multiple sensor arrays with applications to infrasonic systems.