Phone Price Prediction Bigenner Cup

solution to promotion Intermediate
https://signate.jp/competitions/750

Table of Contents

• Competition Description
• Dataset
• Solution Approach
• Results
• Reproduction
• Usage
• Acknowledgments
• License

Competition Description

This competition is open to beginners and registered participants.
If your F1 macro score exceeds 0.462885, you will be recognized as an authorized intermediate participant.

Dataset

0	id	int	Index (used as an identifier)
1	battery_power	int	Total energy a battery can store in one charge (mAh)
2	blue	int	Bluetooth availability (1 if available)
3	clock_speed	float	Clock speed
4	dual_sim	int	Dual SIM support availability (1 if available)
5	fc	int	Front camera megapixels
6	four_g	int	4G support availability (1 if available)
7	int_memory	int	Internal memory (GB)
8	m_dep	float	Mobile depth (cm)
9	mobile_wt	int	Weight
10	n_cores	int	Number of cores
11	pc	int	Primary camera megapixels
12	px_height	int	Pixel resolution height
13	px_width	int	Pixel resolution width
14	ram	int	Random Access Memory (MB)
15	sc_h	int	Screen height of the mobile (cm)
16	sc_w	int	Screen width of the mobile (cm)
17	talk_time	int	Continuous talk time
18	three_g	int	3G support availability (1 if available)
19	touch_screen	int	Touch screen availability
20	wifi	int	Wi-Fi availability (1 if available)
21	price_range	int	Price range:target 0 (low cost), 3 (very high cost)

Solution Approach

preprocess

• minmaxscaler(sk-learn)
• standardscaler(sk-learn)

model

• LightGBM
• NeuralNetwork(Dence)
• KNearestNeighbor

Describe how others can use your code or model for their own predictions. Provide instructions on setting up the environment, installing dependencies, and running the code.

Results

authorized intermediate participant.

Usage

pip install -r requirements.txt

python3 run.py --path [.ini file path] --name [config name]

You have the flexibility to configure each model using the following format: [key] = [type] [value].

Common configurations:

datadir(str): Directory path for the data
model_key(str): Name of the model -->[ nn, knn, lgbm ]
pps(str): Preprocessing method -->[ std, minmax, None ]
load_pretrained(str): Set to either True or False -->[ True, False ]
model_name(str): Name of the pre-trained model to load

Neural Network configuration:

input_shape(int): Shape of the input data
L(int): Number of neurons in the first layer
epoch(int): Maximum number of epochs
batch(int): Batch size
random_state(int): Random seed

KNN configuration:

k(int): Number of nearest points to consider

LGBM configuration:

None

Feel free to adjust these configurations based on your specific needs and model requirements.

License

free.