In the Last time article, I aimed to compress the contents of the securities report to make it easier to read, but in this article, "Which stock buy after all?" I'll try to find out "What should I do?": Rolling_eyes:
Fortunately, COTOHA has an API called ** User Attribute Estimate (β) ** </ font>, so you can anthropomorphize a corporation from a securities report, and ** what kind of legal personality is next. Go with the policy of finding a profit or correlation ** in the year
https://api.ce-cotoha.com/contents/index.html
* Profitable = Assuming that ordinary income is increasing, use [Business status] ⇒ [Analysis of financial status, operating results and cash flow status by management] in the securities report of Nikkei 255 companies. * However, I'm not sure if the stock price will rise just because the ordinary profit has increased **--Find an interesting ** spurious **: rolling_eyes:
The code used for this verification is below https://github.com/ice-github/CoARiJAndCOTOHA
The following is the ** difference (2018-2017) ** of 141 companies that had ordinary income data for 2017 and 2018 in CoARiJ in descending order.
The mean was 22.85, the standard deviation was 425.21, and the median was 3.20. As you can see, the distribution is like that (2σ = 95% ⇒ [-87 billion yen, +83 billion]), so if the difference is positive, it is profitable, and if the difference is negative, it is not profitable **. To go
At first, I tried to collect data by ** user attribute estimation (β) ** </ font> for [analysis of financial condition, business performance and cash flow situation by management]. However, it seems that the result depends on the length of the sentence.
* Since COTOHA cannot receive long sentences, the sentences are divided and thrown to the API, weighted by the number of processed characters, merged and treated as data </ font>
The following are the results of ** 2017 Daikin Industries, Ltd. **
Segment name
Output (million yen)
Year-over-year basis(%)
Air conditioning / refrigerator business
1,548,244
9.6
Chemical business
166,798
19.0
Other businesses
49,125
4.5
total
1,764,168
10.3
(Note) 1 The amount depends on the selling price. 2 The above amount does not include consumption tax. (2) Order status Most of the Group's products are made-to-stock, so the order amount and order backlog are omitted. (3) Sales results The sales results for the current consolidated fiscal year are shown below for each segment.
Segment name
Sales (million yen)
Year-over-year basis(%)
Air conditioning / refrigerator business
2,052,884
11.9
Chemical business
183,147
16.8
Other businesses
54,529
5.2
total
2,290,560
12.1
(Note) 1 Transactions between segments are offset and eliminated. 2 Since the ratio of all customers to the total sales performance is less than 10/100, the description of the sales performance by the other party and the ratio to the total sales performance is omitted. 3 The above amount does not include consumption tax. (Analysis / examination details regarding the status of business results, etc. from the perspective of the management) The contents described below are based on the judgment as of the end of the current consolidated fiscal year. (1) Important accounting policies and estimates The Group's consolidated financial statements are prepared based on accounting standards generally accepted in Japan. The recording of assets, liabilities and net assets at the end of the current consolidated fiscal year and the recording of income and expenses in the current consolidated fiscal year include estimates based on rational standards based on the current situation and past performance. The important accounting policies, etc. for the preparation of consolidated financial statements are as described in "Important matters that are the basis for the preparation of consolidated financial statements." (2) Financial status ① Assets Total assets were 2,488,953 million yen, an increase of 133,855 million yen from the end of the previous consolidated fiscal year. Current assets increased by 77,926 million yen from the end of the previous consolidated fiscal year to 1,237,811 million yen due to an increase in notes and accounts receivable. Fixed assets increased by 55,878 million yen from the end of the previous consolidated fiscal year to 1,252,142 million yen due to an increase due to changes in the market value of investment securities. ② Liability and net assets Liabilities decreased by 54,977 million yen from the end of the previous consolidated fiscal year to 1,165,632 million yen due to a decrease in long-term debt. Net assets increased by 188,712 million yen from the end of the previous consolidated fiscal year to 1,324,321 million yen due to an increase due to the recording of net income attributable to owners of the parent company. As a result, the equity ratio increased from 47.2% at the end of the previous consolidated fiscal year to 52.1%, and the amount of net assets per share increased from 3,802.10 yen at the end of the previous consolidated fiscal year to 4,433.62 yen. (3) Business results ① Sales Sales for the current consolidated fiscal year increased 12.1% from the previous consolidated fiscal year to 2,290,560 million yen. In the air conditioner / refrigerator business, overseas sales were strong, mainly in the Americas, Europe and Asia, and sales increased 11.9% from the previous consolidated fiscal year to 2,052.884 billion yen. In the chemical business, demand for semiconductors and automobiles was strong, and sales increased 16.8% from the previous consolidated fiscal year to 183,147 million yen. In the other business as a whole, sales of hydraulic equipment for industrial machinery and hydraulic equipment for construction machinery and vehicles were firm in the domestic and US markets, and sales increased 5.2% from the previous consolidated fiscal year to 54,529 million yen. .. ② Operating expenses, operating income Cost of sales increased 13.6% from the previous consolidated fiscal year to 1,491,731 million yen. Selling, general and administrative expenses increased 9.0% from the previous consolidated fiscal year to 545,089 million yen. The main factor is the increase in labor costs. As a result of the above, operating income increased 10.0% from the previous consolidated fiscal year to 253,739 million yen. Regarding the operating income of the segment, in the air conditioner / refrigerator business, operating income increased 7.0% from the previous consolidated fiscal year to 223,436 million yen, and in the chemical business, it increased 39.4% from the previous consolidated fiscal year to 255. Operating income was 101 million yen, and other businesses recorded operating income of 4,756 million yen, up 26.8% from the previous consolidated fiscal year. ③ Non-operating income and ordinary income Non-operating income increased by 1,035 million yen from the previous consolidated fiscal year to a plus of 1,279 million yen due to an increase in the amount of investment income recorded by the equity method. Ordinary income increased 10.4% from the previous consolidated fiscal year to 255,019 million yen. (4) Extraordinary gains / losses and net income attributable to owners of the parent company Extraordinary gains / losses decreased by 2,758 million yen from the previous consolidated fiscal year to a minus of 3,162 million yen due to the recording of loss on consolidation of affiliated companies. Net income attributable to owners of the parent increased 22.8% from the previous consolidated fiscal year to 189,051 million yen, partly due to a decrease in corporate taxes due to tax reform in the United States. (4) Cash flow In business activities, income of 43,923 million yen decreased from the previous consolidated fiscal year to 223,740 million yen due to an increase in income taxes and other payments. In investment activities, expenditures decreased by 1,364 million yen from the previous consolidated fiscal year to 127,458 million yen due to a decrease in expenditures due to the acquisition of consolidated subsidiaries. In financial activities, expenditures increased by 20,411 million yen from the previous consolidated fiscal year to 93,954 million yen due to a decrease in short-term borrowings. The increase / decrease in cash and cash equivalents, which is obtained by adding the foreign exchange translation difference to these results, decreased by 39,954 million yen from the end of the previous consolidated fiscal year to 12,933 million yen in cash. It increased. In principle, funds are procured by accumulating retained earnings and focusing on own funds, but if necessary, they are procured by borrowing from financial institutions or corporate bonds. In the current consolidated fiscal year, 45,180 million yen was raised by long-term borrowing from financial institutions and used as part of the investment funds. The trends in cash flow indicators are as follows.
March 2014
Fiscal year ended March 31, 2015
Fiscal year ended March 31, 2016
March 2017
March 2018
Capital adequacy ratio(%)
39.9
45.3
46.3
47.2
52.1
Market value-based capital adequacy ratio (%)
83.9
103.7
112.1
138.8
137.8
Cash Flow to Interest-bearing Debt Ratio (Year)
3.9
4.1
2.7
2.3
2.5
Interest coverage ratio (times)
18.0
16.8
25.9
26.8
20.9
(Note) Capital adequacy ratio: Capital / total assets Market capitalization ratio: Market capitalization / total assets Cash Flow to Interest-bearing Debt Ratio: Interest-bearing Debt / Operating Cash Flow Interest coverage ratio: operating cash flow / interest payment
age: [('30-39 years old', 0.25980392156862747), ('20-29-year-old', 0.05043859649122807), ('40-49 years old', 0.04863261093911249)]
civilstatus: [('married', 0.41885964912280704), ('Unmarried', 0.18601651186790505)]
earnings[('3M-5M', 0.05985552115583075), ('8M-10M', 0.02631578947368421)]
gender[('male', 0.6479618163054695)]
habit[('SMOKING', 0.19272445820433437)]
hobby[('COOKING', 0.9999999999999999), ('INTERNET', 0.8316563467492261), ('SPORT', 0.7804437564499483), ('FISHING', 0.7365841073271413), ('MOVIE', 0.6598297213622292), ('FORTUNE', 0.4854231166150671), ('GYM', 0.43421052631578955), ('GAMBLE', 0.39138286893704854), ('TRAVEL', 0.3117905056759546), ('TVGAME', 0.2502579979360165), ('COLLECTION', 0.21413828689370484), ('CAMERA', 0.1531217750257998), ('TVDRAMA', 0.14641382868937047), ('SHOPPING', 0.11119711042311661), ('ANIMAL', 0.04863261093911249), ('MUSIC', 0.02889576883384933), ('PAINT', 0.027089783281733747), ('TVCOMMEDY', 0.025799793601651185), ('RAILWAY', 0.024638802889576882), ('SPORTWATCHING', 0.024638802889576882), ('IDOL', 0.016898864809081527)]
kind_of_business[('Commercial', 0.03005675954592363)]
kind_of_occupation[('R & D position', 0.02631578947368421)]
location[('Kanto', 0.1059081527347781), ('Kinki', 0.0890092879256966)]
moving[('CAR', 0.46271929824561403), ('WALKING', 0.4300825593395252), ('BUS', 0.04321465428276573), ('RAILWAY', 0.04321465428276573), ('OTHER', 0.028379772961816305), ('NO', 0.024638802889576882)]
occupation[('employee', 0.5687564499484005)]
position[]
age: [('30-39 years old', 0.18653250773993807), ('20-29-year-old', 0.11351909184726522)]
civilstatus: [('married', 0.4273735810113519), ('Unmarried', 0.34300825593395257)]
earnings[('3M-5M', 0.18098555211558306)]
gender[('male', 0.7492260061919505)]
habit[('SMOKING', 0.2043343653250774)]
hobby[('COOKING', 0.935500515995872), ('INTERNET', 0.8125644994840041), ('MOVIE', 0.753482972136223), ('SPORT', 0.7073013415892674), ('FISHING', 0.5370227038183695), ('GYM', 0.42389060887512897), ('TVGAME', 0.4162796697626419), ('TRAVEL', 0.4057017543859649), ('COLLECTION', 0.384029927760578), ('FORTUNE', 0.3635190918472653), ('TVDRAMA', 0.26186790505675955), ('CAMERA', 0.1891124871001032), ('GAMBLE', 0.18537151702786378), ('SHOPPING',
0.1660216718266254), ('PAINT', 0.11816305469556243), ('IDOL', 0.06346749226006192), ('MUSIC', 0.058565531475748195), ('GOURMET', 0.05598555211558308), ('STUDY', 0.05379256965944272), ('SPORTWATCHING', 0.018962848297213623)]
kind_of_business[]
kind_of_occupation[]
location[('Kanto', 0.11880804953560371), ('Kinki', 0.10307017543859648)]
moving[('WALKING', 0.256062951496388), ('CAR', 0.21865325077399383), ('BUS', 0.062306501547987614), ('NO', 0.018962848297213623)]
occupation[('employee', 0.4548503611971104)]
position[]
age: [('30-39 years old', 0.18988648090815274), ('20-29-year-old', 0.0957172342621259), ('40-49 years old', 0.09429824561403509)]
civilstatus: [('married', 0.43021155830753355), ('Unmarried', 0.2866357069143447)]
earnings[('3M-5M', 0.2818627450980392)]
gender[('male', 0.814499484004128)]
habit[('SMOKING', 0.1715686274509804)]
hobby[('INTERNET', 0.8209494324045408), ('COOKING', 0.811016511867905), ('MOVIE', 0.7192982456140351), ('FISHING', 0.7155572755417956), ('SPORT', 0.6235810113519092), ('TRAVEL', 0.5672084623323014), ('GYM', 0.4557533539731683), ('TVGAME', 0.3753869969040248), ('COLLECTION', 0.3668730650154799), ('CAMERA', 0.36119711042311664), ('TVDRAMA', 0.26509287925696595), ('STUDY', 0.18962848297213622), ('FORTUNE', 0.18537151702786378), ('MUSIC', 0.0957172342621259), ('SHOPPING', 0.0957172342621259), ('GAMBLE', 0.09352425180598556), ('IDOL', 0.09352425180598556), ('PAINT', 0.06953044375644994)]
kind_of_business[]
kind_of_occupation[]
location[('Kinki', 0.09429824561403509), ('Kanto', 0.06953044375644994)]
moving[('WALKING', 0.19220846233230132), ('CAR', 0.18330753353973167)]
occupation[('employee', 0.4707172342621259)]
position[]
age: [('30-39 years old', 0.37100103199174406), ('20-29-year-old', 0.12396800825593396)]
civilstatus: [('married', 0.5058049535603716), ('Unmarried', 0.36661506707946334)]
earnings[('3M-5M', 0.36119711042311664), ('-1M', 0.12319401444788441)]
gender[('male', 0.7381320949432404)]
habit[('SMOKING', 0.1385448916408669)]
hobby[('COOKING', 0.8817079463364293), ('MOVIE', 0.7549019607843138), ('INTERNET', 0.7381320949432404), ('FISHING', 0.628482972136223), ('SPORT', 0.6273219814241486), ('GYM', 0.612358101135191), ('TRAVEL', 0.3691950464396285), ('TVGAME', 0.36906604747162025), ('STUDY', 0.3618421052631579), ('CAMERA', 0.26986584107327144), ('TVDRAMA', 0.26986584107327144), ('COLLECTION', 0.26044891640866874), ('FORTUNE', 0.2465170278637771), ('MUSIC', 0.24587203302373584), ('PAINT', 0.24316305469556243), ('SHOPPING', 0.23787409700722395), ('GAMBLE', 0.11996904024767802), ('SPORTWATCHING', 0.018962848297213623)]
kind_of_business[]
kind_of_occupation[]
location[('Kanto', 0.12319401444788441), ('Kinki', 0.11829205366357069)]
moving[('WALKING', 0.26986584107327144), ('CAR', 0.12332301341589268), ('NO', 0.018962848297213623)]
occupation[('employee', 0.4931630546955625)]
position[]
When separated by 250 characters ** A married man in his 30s with an income of '3-5M (3 million yen or more and 5 million yen or less)' smokes cigarettes Cooking, Internet and sports < My hobby is / font, and I'm an office worker who lives in the Kanto region and commute by car </ font> **.
Divided by 1000 characters ** A married man in his 30s with an income of '3-5M (3 million yen or more and 5 million yen or less)' smokes cigarettes Cooking, movies and the Internet </ font> > Is my hobby, and I will be an office worker who lives in the Kanto region and walks to work </ font> **
Since the order of the items is almost the same, I will ignore the numerical value of each item and I will output the result separated by 1000 characters (to reduce the number of API calls): upside_down:
Of the 141 companies, the ** recurring profit difference ** from 2017 to 2018 was 75 positive and 66 negative.
** Bin count the largest elements of each attribute (Hobby = top 3 hobbies) and compare the distribution ** Since the number of data is 75 and 66 respectively, divide the number by 75 and 66 for easy comparison
* Is this method the same as applying a black box hash function from the viewpoint of feature engineering? </ font>
There is almost no difference
Marriage rate is lower for those who are slightly profitable, but there is almost no difference
** Those who are profitable have a low annual income **
There was no difference and both were men
The smoking rate is slightly higher for those who are profitable, but there is almost no difference. * Habit should have other items, but I don't know what they are. </ Font>
** Those who are profitable have high FORTUNE (fortune-telling) and low SPORT (sports) **
There is almost no difference
There is almost no difference
** There are few Kinki and many Kanto people who are profitable **
** Those who are profitable move a lot by car ** (Slightly less walking)
There is almost no difference * Occupation should have other items, but I don't know what they are. </ Font>
Those who are profitable have a few chiefs
--For those who are profitable ** Annual income is low (mostly 0 to 3 million yen) ** --For those who are profitable ** Hobbies are high for FORTUNE (fortune-telling) and low for SPORT (sports) ** --For those who are profitable ** There are few places in Kinki and many in Kanto **
In this article, COTOHA's ** User Attribute Estimate (β) API ** </ font> is used in the securities report [Management's financial position, operating results and cash flow status. Analysis] was applied to explore what attributes correlate with the increase in ordinary income for the next year (result ↓)
I tried it because COTOHA can do this, but originally ** user attribute estimation (β) ** </ font> is user support email and phone content. You would use it for transcribed data, but you wouldn't expect it to be used for data in such a domain: sweat_smile: ** It's interesting to try quickly and get the result, "What happens if you throw this kind of data for the time being?" **
* It was quite difficult to use it only 1000 times a day, so it would be nice to double it (2000 times) ... </ font>
The companies that searched for are as follows
--6479: MinebeaMitsumi Co., Ltd. - Profit decreased by 29.8% </ font> --7202: Isuzu Motors Ltd. - Ordinary income decreased by 22.9% </ font> --7211: Mitsubishi Motors Corporation - Ordinary income is in the red of 2.6 billion yen </ font> ―― 7270: SUBARU Corporation - Profit decreased by 4.8% </ font> --4902: Konica Minolta Co., Ltd. - Profit decreased by 93.7% </ font> --82333: Takashimaya Co., Ltd. - Ordinary income decreased by 15.2% </ font> ---8830: Sumitomo Realty & Development Co., Ltd. - Ordinary income increased by 6.2% </ font>
* If you did not find the item of ordinary income in the third quarter financial results, you are looking at the item of operating income or (net) income </ font>
Recommended Posts