Task

An loT system containing three sensors placed in three physical locations with variable environmental conditions is

used to identify motion in the proximity of the sensors$.$ Sensor readings were collected regularly for a certain period.

Data collected is stored in the file data.cvs attached to this assignment.

The data file contains the following data columns:

ts $-$ time stamp

device $-$ sensor ID

co $-$ environmental carbon monoxide

humidity $-$ environmental humidity

light $-$ boolean to indicate if light is detected

gas $-$ environmental gas

motion $-$ Boolean to indicate if motion is detected

smoke $-$ environmental smoke

temp $-$ environmental temperature

Using this dataset, design and implement a convolutional neural network $($CNN$)$ to model the "motion" feature $($the

model developed should predict the feature "motion"$).$ Optimise the model $($its hyperparameters$)$ and analyse its

performance.

In implementing the CNN$,$ use Keras or another preferred Python based Al tool.

Present the design, implementation, and optimisation process, together with the analysis of the performance in a

technical report.

The report should focus on the practical achievements with diagrams, plots, tables, numerical results. Include only

the necessary text to clearly explain the practical work $($avoid extensively long text$).$

The report should contain only the key points about the code implementation and testing. The full code should be

submitted separately $($see below$).$